Modern AI services must continually adapt to newly joined domains, yet delivering high-quality customized models is hampered by label sparsity, domain shifts, and tight budgets. We formulate this challenge as the learning system expansion problem and introduce HaT, an efficient heterogeneity-aware knowledge-transfer framework. HaT first selects a small set of high-quality source models with minimal overhead, and then fuses their imperfect predictions through a sample-wise attention mixer. Later, it adaptively distills the fused knowledge into target models via a knowledge dictionary. Extensive experiments on different tasks and modalities show that HaT outperforms state-of-the-art baselines by up to 16.5% accuracy, and saves 31.1% training time and up to 93.0% traffic.

The analog-to-digital converter (ADC) in a radio frequency (RF) front-end and digital signal processing (DSP) are significant sources of energy consumption, particularly in low-power systems like LoRa. To reliably demodulate weak signals, current systems rely on heavy oversampling - often 8x the signal bandwidth - which imposes a substantial and persistent oversampling tax on the analog front-end and DSP. This paper investigates if this tax can be mitigated by adapting techniques from image and video super resolution. We propose RF Super Resolution, a lightweight, real-time neural upscaler for RF signals. Our approach pairs an efficient digital interpolation algorithm with a shallow four-layer CNN. The neural network is trained to learn and correct the residual artifacts introduced by the digital upsampling and noise, effectively mimicking the output of a high-rate analog ADC and denoising filter. We validate our system on a large-scale, over-the-air LoRa study. Our results show that RF-SR, given a 2× Nyquist input (250 kHz), restores demodulation performance of a native 8× oversampled (2 MHz) system at half its sampling rate (1 MHz). This effectively removes the analog oversampling requirement, and provides an additional 1.25 dB SNR gain over the oversampled baseline, making it an efficient and effective signal enhancer suitable for gateway integration or post training quantized deployment at the end-node.

Modern autonomous vehicles face growing cybersecurity risks, especially from action space attacks that directly target vehicle actuators. This article systematically evaluates the resilience of three representative Autonomous Driving (AD) architectures, including modular, end-to-end, and feature-fused agents, against few-shot action space attacks crafted via deep reinforcement learning under a black-box setting. The adversary perturbs the vehicle’s lateral control only during safety-critical moments, using either a camera or an inertial measurement unit. Our results reveal distinct vulnerabilities and behavioral patterns across AD architectures, which underscore the necessity for adaptive and robust defense strategies. However, existing adversarial training defense methods show limitations of overfitting and reliance on attack knowledge. To address these limitations, we propose a learning-based Path Correction System (PCS) that integrates traditional feedback control with an adversarially trained correction loop. The correction loop is selectively activated by a kinematic model-based attack detector to counteract abnormal control deviations. Evaluation experiments show that PCS reduces path-tracking deviation by 78% when the system is under attack.

A modern data center is supported by an Internet of Assets (IoA), a specialized Internet of Things (IoT), where the sim-ready assets encompass physical properties and connected data streams for passive data collection and proactive simulation. A digital twin (DT) is a virtual replica of the IoA system with a proper level of abstraction, integrating asset-level dynamics models to simulate system-wide behavior, prototype designs, and conduct what-if analysis. However, creating high-fidelity DTs is hindered by the manual effort needed to encode complex geometric layouts and domain-specific design constraints. While large language models (LLMs) offer automation potential, existing methods struggle to generate geometrically plausible and functionally valid DT scenes due to limited domain integration. To bridge the gaps, we propose ChatDC, a conversational system that leverages LLMs to automate data center digital twin generation through a S egment- G enerate- O ptimize (SGO) workflow. ChatDC integrates domain knowledge via a dedicated code library named DCBuild and employs SGO to decompose user prompts, generate initial structures, and optimize layouts in compliance with data center design constraints. Evaluation shows that ChatDC outperforms other baselines with 98% success rate on scratch generation tasks and reduces the average makespan by 10×. Ablation study reveals that the SGO design increases the generation success rate by 65% at most. Furthermore, computational fluid dynamics simulations validate the physical plausibility, confirming the readiness of generated DTs for real-world analysis.

Existing autonomous vehicles (AVs) utilize neither cloud computing for execution of their deep learning-based driving tasks due to the long vehicle-to-cloud communication latency, nor solar energy to offset the energy usage of car-borne computing. They are in general equipped with the resource-constrained edge computing devices which may be unable to execute the compute-intensive deep learning models in real time. The increasing data transmission speed of the commercial mobile networks sheds light upon the feasibility of using the cloud computing for autonomous driving, which is demonstrated by our city-scale real-world measurements over the fifth generation (5G) mobile networks. Moroever, the cost and form factor declines of solar harvesting systems make the quest of integrating them with AVs for improving carbon efficiency more promising. In this paper, we present the design and implementation of ECSeg, an edge-cloud switched low-carbon image segmentation system for AVs equipped with roof-mounted solar panels. ECSeg dynamically switches between edge and cloud processing to execute semantic segmentation models in real time while aiming to decarbonize AV computing by maximizing the use of harvested solar energy. The switching decision is challenging due to the complex interdependencies among various factors, including dynamic wireless channel conditions, vehicle motion, visual scene changes, and available renewable energy. To tackle this, ECSeg employs deep reinforcement learning to learn an optimal switching policy. Extensive evaluations based on real-world experiments and trace-driven simulations demonstrate that ECSeg outperforms six baseline approaches, achieving 98.8% reduction in computing’s carbon emission, while maintaining high image segmentation accuracy, compared with our earlier design without integrating the solar panel.

Time series is an important form of data generated by Internet-of-Things (IoT) devices. Closed-form continuous-time (CFC) neural networks offer superior expressivity for modeling time series data compared with recurrent neural networks. Additionally, their lower training and inference overhead make them well-suited for deployment on microcontroller-based IoT devices. This paper introduces FedCFC, an on-device federated learning network that operates based on the CFC models distributed across IoT devices. FedCFC incorporates a novel and communication-efficient aggregation strategy designed to mitigate the effects of class distribution imbalances across the participating IoT devices’ training data. The strategy is designed based on a new property of CFC identified in this paper, i.e., the insensitivity of a sub-model of CFC with respect to training data’s class distribution shift. Extensive evaluation with multiple time series datasets demonstrates that FedCFC attains comparable or superior accuracy while achieving a 7.6× to 11× reduction in communication overhead compared with recent federated learning approaches designed to address the class distribution skew problem. Furthermore, deployments of FedCFC on four IoT platforms highlight its suitability for resource-constrained devices with as little as 256 kB of memory or even less.

High-definition (HD) maps provide precise environmental information essential for prediction and planning in autonomous driving (AD) systems. Due to the high cost of labeling and maintenance, recent research has turned to online HD map construction using onboard sensor data, offering wider coverage and more timely updates for autonomous vehicles (AVs). However, the robustness of online map construction under adversarial conditions remains underexplored. In this paper, we present a systematic vulnerability analysis of online map construction models, which reveals that these models exhibit an inherent bias toward predicting symmetric road structures. In asymmetric scenes like forks or merges, this bias often causes the model to mistakenly predict a straight boundary that mirrors the opposite side. We demonstrate that this vulnerability persists in the real-world and can be reliably triggered by obstruction or targeted interference. Leveraging this vulnerability, we propose a novel two-stage attack framework capable of manipulating online constructed maps. First, our method identifies vulnerable asymmetric scenes along the victim AV's potential route. Then, we optimize the location and pattern of camera-blinding attacks and adversarial patch attacks. Evaluations on a public AD dataset demonstrate that our attacks can degrade mapping accuracy by up to 9.9% in average precision, render up to 44% of targeted routes unreachable, and increase unsafe planned trajectory rates—colliding with real-world road boundaries—by up to 27%. These attacks are also validated on a real-world testbed vehicle. We further analyze root causes of the symmetry bias, attributing them to training data imbalance, model architecture, and map element representation. Based on these findings, we propose asymmetric data fine-tuning as a targeted defense, which significantly improves model robustness. To the best of our knowledge, this study presents the first vulnerability assessment of online map construction models and introduces the first digital and physical attack against them.

Stereo vision plays a critical role in enabling depth perception for drones, supporting navigation and obstacle avoidance in complex environments. However, the robustness and security of stereo vision systems remain largely underexplored. In this paper, we propose Rolling in the Deep (RiD), a novel physical attack that exploits the rolling shutter effect (RSE) to inject imperceptible, structured perturbations into stereo image pairs. We analyze RSE formation in binocular camera setups and show how RSE-based perturbations can degrade deep learning-based stereo matching by exploiting model vulnerabilities and sensor misalignments, resulting in incorrect depth estimation. Preliminary results show the feasibility of RiD under realistic stereo configurations, revealing a new class of threats to drone perception systems.

Adversarial attacks with real objects or lasers on car-borne LiDAR-based object detection are concerning. The existing defense approaches are often designed to address specific attacks and short of considering adaptive attackers who may adapt based on all available information about the deployed defense to maximize attack effect. This paper proposes Hyper3Def, a new defense for the function of detecting vehicle objects, which uses a Hypernet to generate an ensemble of multiple new detection models when needed at run time. The detection results of these models are fused to give the final result. As a dynamic defense, Hyper3Def revokes an important basis of the adaptive attack, i.e., the object detection model is needed to plan effective adversarial perturbations. Evaluation based on open data and real-world experiments with embedded system implementation show that, when confronting adaptive attacks, Hyper3Def outperforms various baseline defenses including the adversarial training, which is often cited as the state of the art.

Stochastic differential equation networks (SDENets), a subset of continuous-time neural networks, offer the natural ability to model continuous time-series data with greater expressivity than discrete-time neural networks. However, SDENets face challenges related to stability and high computational overhead. In this paper, we introduce SE-SDENet, a stable and efficient variant of SDENet. Leveraging the inherent capability of SDENets to model randomness in time-series data, we establish a theoretical framework that ensures SE-SDENet's stability during training by regulating the dynamics of each neuron. Additionally, we propose an efficient training and inference framework that enables SE-SDENet to achieve low forward-pass complexity and to dynamically adjust its complexity at run-time. Evaluation with four datasets and four edge devices demonstrates that SE-SDENet achieves a 6.6x higher throughput than the solver-based SDENet and exhibits improved stability in handling noisy data and long-term predictions. A validation on a robot vehicle shows that SE-SDENet can dynamically adjust its complexity at run-time to meet varying resource constraints.

Deep neural networks (DNNs) have achieved notable success in mobile vision tasks, yet they show vulnerability to adversarial attacks. When carefully crafted perturbations are introduced, these models can be easily misled into wrong classifications, posing significant risks for safety-critical mobile systems like autonomous vehicles. Although various defense strategies, both static and dynamic, have been proposed, many fail to address adaptive attacks or overlook the resource constraints of mobile systems. To address these limitations, in this paper, we present GuSoDrop, a lightweight dynamic defense framework that applies stochastic layer dropping. GuSoDrop leverages randomness to counteract adaptive attacks while selectively dropping less important layers to reduce computation overhead. Our preliminary evaluation shows that GuSoDrop outperforms state-of-the-art defense methods against different adaptive attacks and improves efficiency in reducing computational overhead.

Camera-based object detection excels but remains vulnerable to adversarial attacks that suppress target detection (object-hiding attacks). Here, we propose PaSED, a Parameterized Stochastic Ensemble Defense, which leverages HyperNetworks to enable rapid and diverse updates for detection models in the ensemble. At its core, we introduce functional diversity to enhance the defense robustness. It adapts each generation process to the input image preprocessing parameterized by HyperNetworks' random noise input. In our preliminary evaluations against physically deployed attacks, PaSED outperforms five baseline defenses without requiring attack knowledge. It recovers attacked objects in 92% and 98% of frames in the indoor and outdoor testbeds, respectively.

Existing autonomous vehicles have not utilized the cloud computing for execution of their deep learning-based driving tasks due to the long vehicle-to-cloud communication latency. The increasing data transmission speed of the commercial mobile networks sheds light upon the feasibility of using the cloud computing for autonomous driving. In this demo, we introduce the design and implementation of ECSeg, an edge-cloud switched image segmentation system that dynamically selects between the edge and cloud to execute deep learning-based semantic segmentation models. This enables realtime understanding of a vehicle's visual scenes while adapting to dynamic wireless conditions and changing environments.

To maintain driving safety, the execution of neural network-based autonomous driving pipelines must meet the dynamic deadlines in response to the changing environment and vehicle’s velocity. To this end, this article proposes a real-time neural architecture adaptation approach, called TimelyNet, which uses a supernet to replace the most compute-intensive neural network module in an existing end-to-end autonomous driving pipeline. From the supernet, TimelyNet samples subnets with varying inference latency levels to meet the dynamic deadlines during run-time driving without fine-tuning. Specifically, TimelyNet employs a one-shot prediction method that jointly uses a lookup table and an invertible neural network to periodically determine the optimal hyperparameters of a subnet to meet its execution deadline while achieving the highest possible accuracy. The lookup table stores multiple subnet architectures with different latencies, while the invertible neural network models the distribution of the optimal subnet architecture given the latency. Extensive evaluation based on hardware-in-the-loop CARLA simulations shows that TimelyNet-integrated driving pipelines achieve the best driving safety, characterized by the lowest wrong-lane driving rate and zero collisions, compared with several baselines, including the state-of-the-art driving pipelines.

Deep neural networks have advanced the perception and decision-making functions of smart embedded systems, such as car-borne driver assistance. Deploying these embedded neural networks often faces two challenges: (i) security vulnerabilities to adversarial examples that can be deployed in the perceived physical environment; (ii) limited computational resources coupled with dynamic conditions that necessitate real-time adaptation of model execution. However, these two challenges are often addressed separately in existing research. This article presents LeapNet, which aims to address both challenges simultaneously. It comprises two versions: LeapNet-1 and LeapNet-2. LeapNet-1 employs dynamic layer routing to counteract adaptive adversarial-example attacks and reduce computational redundancy. Building upon LeapNet-1, LeapNet-2 further adapts its layer routing configurations in real time to meet the frame processing rate requirements under dynamic conditions while maintaining defense performance. Extensive experiments on various representative datasets, neural network models, and adaptive attacks demonstrate the superiority of LeapNet over existing defense methods. On-road tests with a real-time car-borne traffic sign recognition system validate its effectiveness in maintaining frame processing rate under dynamic conditions.

Face authentication (FA) schemes are widely adopted in smart homes nowadays. However, existing FA systems for smart appliances are commonly camera-based and hence experience performance degradation in poor illumination conditions. Mainstream FA systems based on radio frequency require dedicated hardware that is inaccessible to many appliances. In this paper, we propose an acoustic signals-based FA scheme that extracts acoustic signal features associated with facial 3D geometries to achieve FA named SoundFace . This scheme can be widely deployed on most appliances in home environments. We propose a novel two-stage locating approach based on acoustic sensing to capture the signal variation of the user’s face and separate the face region echoes from multipath interferences in the distance dimension. To obtain distinguishable facial features, we design a Convolutional Neural Network (CNN)-based feature extractor. In addition, the acoustic signal is highly susceptible to different changes in practical authentication. To overcome it, we utilize a transfer learning technique with little training overhead to enable SoundFace resilient to various authentication changes. Extensive evaluations demonstrate that SoundFace achieves an average true authentication rate of over 96.2% and an equal error rate of 4.2%, and it is robust to various real-world settings.

The 3D-printed mesh screen with dense penetrating pores is a new structure for massive manufacturing of molded pulp package products. However, some of the pores may be clogged by the printing material powder during the printing process. Such defects negatively affect the quality of the pulp packages produced using the mesh screen mold. To pinpoint the defects, we design a model-based robotic visual sensing system, called RoboCam, which uses a robotic arm to carry a high-resolution camera for full inspection of a mold consisting of joined mesh screens. To inspect the entire mold, RoboCam plans the camera poses to capture multiple images of the mold and render synthesized images as references for identifying the clogged pores. In particular, we propose novel designs to rectify the inherent run-time pose errors of the robotic system for ensuring the reference quality and to accelerate the reference rendering for reducing inspection latency. Extensive evaluation shows that RoboCam’s design outperforms various baselines, including three existing computer vision and convolution neural network-based inspection systems. RoboCam achieves a recall rate of 94.95% within 528 seconds latency for inspecting an entire mold with 13,000 designed pores.

The new real-time interactive services, such as virtual and augmented reality, demand significantly higher network bandwidth and quality, which the traditional centralized cloud struggles to meet. In addition, centralized optimization management becomes inefficient as the scale of the scene continues to expand. In response, edge cloud systems have emerged, but distributed geographic locations, burstable billing business models, and large numbers of servers in large-scale scenarios pose new challenges for resource management. In this article, we proposeGeoCC, a novel strategy to save bandwidth overhead in burstable billing edge cloud systems.GeoCCaddresses challenges through a dual approach. First, a geography-aware graph construction and partitioning algorithm is used to organize server resources, and a large number of servers are reasonably divided into multiple server pools for parallel processing. Second, it introduces an enhanced burstable billing optimization mechanism that considers contextual factors and adaptive bandwidth capacity. Experiments based on real data from an edge cloud operator demonstrate the effectiveness ofGeoCC. Compared with the baseline,GeoCCcan effectively reduce bandwidth peaks, decreasing bandwidth costs by an average of 28.30% and up to 81.83% at the 95th percentile billing.

This paper considers the carbon-aware data center (DC) capacity provisioning problem under uncertain green energy availability and computing demand. To address it, accurate carbon emissions estimation and robust capacity provisioning are necessary. Existing studies mainly consider the carbon footprint of the computing system and merely consider that of the physical facilities, which also contribute significant carbon emissions. Furthermore, their capacity provisioning is neither uncertainty-aware nor adaptive to the dynamic computing demand. To bridge these gaps, we propose an adaptive capacity provisioning framework based on the physics-informed digital twin. We design the digital twin to holistically capture a DC's operational carbon footprint, including both the computing system and the physical facilities. The digital twin is differentiable and established with a collection of physics-informed learnable models that are learned with online operational data. We further address the challenge of capacity provisioning under uncertainties by designing a shrinking horizon model predictive control. The designed capacity planner updates its estimation of future computing demand based on the observable computing system states. At each capacity provisioning round, we solve the capacity provisioning problem using a gradient-based optimization technique with the gradient provided by the digital twin. We extensively evaluate our approach usingrealoperational data from a large-scale production data center. First, our digital twin accurately predicts holistic DC energy usage with a relative absolute error of less than 5%, which is accurate according to the industrial rule of thumb. Second, we show that our solution is comparable to the oracle solution with perfect knowledge about all uncertainties, outperforming the state-of-the-art Predict-then-Plan approach significantly in terms of SLO violation reduction. Furthermore, our approach reduces carbon footprint by 27% compared with the over-provisioning scheme currently adopted by the industry.

Autonomous driving systems rely on precise trajectory prediction for safe and efficient motion planning. Despite considerable efforts to enhance prediction accuracy, inherent uncertainties persist due to data noise and incomplete observations. Many strategies entail formalizing prediction outcomes into distributions and utilizing variance to represent uncertainty. However, our experimental investigation reveals that existing trajectory prediction models yield unreliable uncertainty estimates, necessitating additional customized calibration processes. On the other hand, directly applying current calibration techniques to prediction outputs may yield sub-optimal results due to using a universal scaler for all predictions and neglecting informative data cues. In this paper, we propose Customized Calibration Temperature with Regularizer (CCTR), a generic framework that calibrates the output distribution. Specifically, CCTR 1) employs a calibration-based regularizer to align output variance with the discrepancy between prediction and ground truth and 2) generates a tailor-made temperature scaler for each prediction using a post-processing network guided by context and historical information. Extensive evaluation involving multiple prediction and planning methods demonstrates the superiority of CCTR over existing calibration algorithms and uncertainty-aware methods, with significant improvements of 11%-22% in calibration quality and 17%-46% in motion planning.

Closed-form continuous-time (CFC) neural networks have superior expressivity in modeling time series data compared with recurrent neural networks. CFC’s lower training and inference overheads also make it appealing for microcontroller-based platforms. This paper proposes FedCFC, which advances CFC from the centralized learning setting to the federated learning paradigm. FedCFC features a novel and communication-efficient aggregation strategy to address the problem of class distribution skews across clients’ training data. The strategy is designed based on a new empirical property of CFC identified in this paper, i.e., involatility of a sub-network of CFC with respect to training data’s class distribution. Extensive evaluation based on multiple time series datasets shows that FedCFC achieves higher or similar accuracy with 7.6× to 11× reduction in communication overhead, compared with recent federated learning approaches designed to address the class distribution skew problem. Implementations of FedCFC on four microcontroller platforms show its portability to low-end computing devices with 256kB memory and even less.

Increasing supply air temperatures is a rule-of-thumb approach to reduce cooling energy usage of data centers (DCs). However, colocation DCs are short of incentive programs to move tenants from the current over-cooling strategy despite the expanding allowable temperature ranges of the computing equipment. This paper considers an essential incentive mechanism, in which the DC operator offers monetary incentives to offset tenants’ electricity payments. We propose an encoder-embedded multi-agent reinforcement learning solution to let the operator agent and tenant agents collaboratively find their policies for deciding the incentives and supply air temperatures, respectively, which are coupled in determining the DC’s total cooling power usage. The solution does not require the cooling power model, which is complex and in general unavailable in practice. Moreover, as each tenant agent learns in the other tenants’ latent state spaces defined by their pre-trained variational autoencoders, only encoded tenants’ states are exchanged, thereby mitigating information leakage concerns. Extensive trace-driven evaluation and comparison with three baselines show that our solution effectively incentivizes tenants to move from the over-cooling strategy and achieves substantial cooling power savings.

Camera-based computer vision is essential to autonomous vehicle's perception. This paper presents an attack that uses light-emitting diodes and exploits the camera's rolling shutter effect to create adversarial stripes in the captured images to mislead traffic sign recognition. The attack is stealthy because the stripes on the traffic sign are invisible to human. For the attack to be threatening, the recognition results need to be stable over consecutive image frames. To achieve this, we design and implement GhostStripe, an attack system that controls the timing of the modulated light emission to adapt to camera operations and victim vehicle movements. Evaluated on real testbeds, GhostStripe can stably spoof the traffic sign recognition results for up to 94% of frames to a wrong class when the victim vehicle passes the road section. In reality, such attack effect may fool victim vehicles into life-threatening incidents. We discuss the countermeasures at the levels of camera sensor, perception model, and autonomous driving system.

Existing autonomous vehicles have not utilized the cloud computing for execution of their deep learning-based driving tasks due to the long vehicle-to-cloud communication latency. Meanwhile, the vehicles are in general equipped with the resource-constrained edge computing devices which may be unable to execute the compute-intensive deep learning models in real time. The increasing data transmission speed of the commercial mobile networks sheds light upon the feasibility of using the cloud computing for autonomous driving. Our city-scale real-world measurements show that the vehicles can partially use the cloud computing via the fifth generation (5G) mobile network with the low data transmission latency. In this paper, we present the design and implementation of ECSeg, an edge-cloud switched image segmentation system that dynamically switches between the edge and cloud for executing the deep learning-based semantic segmentation models to understand the vehicle's visual scenes in real time. The switching decision-making is challenging due to the intricate interdependencies among various factors including the dynamic wireless channel condition, vehicle's movement and visual scene change. To this end, we employ deep reinforcement learning to learn an optimal switching policy. Extensive evaluation based on both real-world experiments and trace-driven simulations demonstrates that ECSeg achieves superior image segmentation accuracy for autonomous vehicles, compared with four baseline approaches.

Camera-based computer vision is crucial for autonomous vehicle perception. We demonstrate GhostStripe [5], an attack system that uses light-emitting diodes and exploits the camera's rolling shutter effect to generate adversarial stripes that are invisible to humans while misleading traffic sign recognition. To maintain stable attack effectiveness, GhostStripe controls the timing of the modulated light emission, adapting to both the camera's framing operation and the movement of the victim vehicle. Evaluated on real testbeds, GhostStripe can stably spoof traffic sign recognition results for up to 97% of frames to a wrong class when the victim vehicle passes the road section.

Visual sensing has been widely adopted for quality inspection in production processes. This article presents the design and implementation of a smart collaborative camera system, called BubCam , for automated quality inspection of manufactured ink bags in Hewlett-Packard (HP) Inc.’s factories. Specifically, BubCam estimates the volume of air bubbles in an ink bag, which may affect the printing quality. The design of BubCam faces challenges due to the dynamic ambient light reflection, motion blur effect, and data labeling difficulty. As a starting point, we design a single-camera system that leverages various deep learning (DL)-based image segmentation and depth fusion techniques. New data labeling and training approaches are proposed to utilize prior knowledge of the production system for training the segmentation model with a small dataset. Then, we design a multi-camera system that additionally deploys multiple wireless cameras to achieve better accuracy due to multi-view sensing. To save power of the wireless cameras, we formulate a configuration adaptation problem and develop the single-agent and multi-agent deep reinforcement learning (DRL)-based solutions to adjust each wireless camera’s operation mode and frame rate in response to the changes of presence of air bubbles and light reflection. The multi-agent DRL approach aims to reduce the retraining costs during the production line reconfiguration process by only retraining the DRL agents for the newly added cameras and the existing cameras with changed positions. Extensive evaluation on a lab testbed and real factory trial shows that BubCam outperforms six baseline solutions including the current manual inspection and existing bubble detection and camera configuration adaptation approaches. In particular, BubCam achieves 1.3x accuracy improvement and 300x latency reduction compared with the manual inspection approach.

Deep reinforcement learning (DRL) has shown good performance in tackling Markov decision process (MDP) problems. As DRL optimizes a long-term reward, it is a promising approach to improving the energy efficiency of data-center cooling. However, enforcement of thermal safety constraints during DRL’s state exploration is a main challenge. The widely adopted reward-shaping approach adds negative reward when the exploratory action results in unsafety. Thus, it needs to experience sufficient unsafe states before it learns how to prevent unsafety. In this article, we propose a safety-aware DRL framework for data-center cooling control. It applies offline imitation learning and online post-hoc rectification to holistically prevent thermal unsafety during online DRL. In particular, the post-hoc rectification searches for the minimum modification to the DRL-recommended action such that the rectified action will not result in unsafety. The rectification is designed based on a thermal state transition model that is fitted using historical safe operation traces and able to extrapolate the transitions to unsafe states explored by DRL. Extensive evaluation for chilled water and direct expansion-cooled data centers in two climate conditions show that our approach saves 18% to 26.6% of total data-center power compared with conventional control and reduces safety violations by 94.5% to 99% compared with reward shaping. We also extend the proposed framework to address data centers with non-uniform temperature distributions for detailed safety considerations. The evaluation shows that our approach saves 14% power usage compared with the PID control while addressing safety compliance during the training.

Voltage stability assessment is essential for maintaining reliable power grid operations. Stability assessment approaches using deep learning address the shortfalls of the traditional time-domain simulation-based approaches caused by increased system complexity. However, deep learning models are shown to be vulnerable to adversarial examples in the field of computer vision. While this vulnerability has been noticed by the power grid cybersecurity research, the domain-specific analysis on the requirements imposed upon effective attack implementation is still lacking. Although these attack requirements are usually reasonable in computer vision tasks, they can be stringent in the context of power grids. In this paper, we conduct a systematic investigation on the attack requirements and credibility of six representative adversarial example attacks based on a voltage stability assessment application for the New England 10-machine 39-bus power system. We show that (1) compromising about half the transmission system buses’ voltage traces is a rule-of-thumb attack requirement; (2) the universal adversarial perturbations regardless of the original clean voltage trajectory possess the same credibility as the widely studied false data injection attacks on power grid state estimation, while the input-specific adversarial perturbations are less credible; (3) the prevailing strong adversarial training thwarts the universal perturbations but fails in defending certain input-specific perturbations. To advance defense to cope with both universal and input-specific adversarial examples, we propose a new approach that simultaneously estimates the predictive uncertainty of any given input of voltage trajectory and thwarts the attacks effectively.

Various deep neural networks (DNNs) including convolutional neural networks (CNNs) and recurrent neural networks (RNNs) have shown appealing performance in various classification tasks. However, due to their large sizes, a single DNN often cannot fit into the memory of resource-constrained smart IoT sensors. This paper presents a DNN splitting framework calledNNFacetthat aims to run a DNN-based classification task on a total of$N$concurrent battery-based sensors observing the same physical process. We begin with determining the importance of all CNN filters or RNN units in learning each class. Then, an optimization problem divides the class set into$N$subsets and assigns them to the sensors, where the important CNN filters or RNN units associated with a class subset form a small model that is deployed to a sensor. Lastly, a multilayer perceptron is trained and deployed to a cloud or edge server, which yields the final classification result based on the low-dimensional features extracted by the sensors using their small models for the same observation. We apply NNFacet to three case studies of voice sensing, vibration sensing, and visual sensing. Extensive evaluation shows that NNFacet outperforms four baseline approaches in terms of system lifetime, latency, and classification accuracy.

Indoor self-localization has become a highly desirable system function for smartphones. The existing systems based on imaging, radio frequency, and geomagnetic sensing may have sub-optimal performance when their limiting factors prevail. In this paper, we present a new indoor simultaneous localization and mapping (SLAM) system that is based on the smartphone's built-in audio hardware and inertial measurement unit (IMU). Our system uses a smartphone's loudspeaker to emit near-inaudible chirps and then the microphone to record the acoustic echoes from the indoor environment. The echoes contain the smartphone's location information with sub-meter granularity. To enable SLAM, we apply contrastive learning to train an echoic location feature (ELF) extractor, such that the loop closures on the smartphone's trajectory can be accurately detected from the associated ELF trace. The detection results effectively regulate the IMU-based trajectory reconstruction. The reconstructed trajectories are used fortrajectory map superimpositionandroom geometry reconstruction. Extensive experiments show that our SLAM achieves median localization errors of$\text{0.1}\,\text{m}$,$\text{0.53}\,\text{m}$, and$\text{0.4}\,\text{m}$in a living room, an office, and a shopping mall, and outperforms both the Wi-Fi and geomagnetic SLAM systems. The room geometry reconstruction achieves up to 4× lower errors compared with the latest echo-based approaches.

Digital Twin (DT) technologies create digital models of physical entities frequently in multimedia forms, which are crucial for concurrent simulation and analysis of real-world systems. In displaying DTs, Holographic-Type Communication (HTC) provides immersive multimedia access for users to interact with Holographic DTs (HDTs) by transmitting holographic data such as Light Field (LF) and other multisensory information. HDT has applications in remote education, work, and social interactions. However, the effective matching of demand and supply between HDT users and providers remains a challenge. To address this issue, we propose a hierarchical architecture that integrates the DT and HTC paradigms. This architecture incorporates a marketplace for HDT services, leveraging a formulated Double Dutch Auction (DDA) mechanism to optimize matching and pricing based on user and provider valuation. Furthermore, We employ an actor-critic-based Deep Reinforcement Learning (DRL) algorithm to train a DDA auctioneer that dynamically adjusts auction clocks during the auction process. As an alternative to the Multi-layer Perceptron (MLP), we experiment with a Deep Simplistic Variational Quantum Circuit (DSVQC) to reduce the number of parameters and enhance performance stability. Our simulations reveal that the proposed learning-based auctioneer achieves 92% optimal social welfare at a 37% auction information exchange cost for an MLP-based actor and 99% optimal social welfare at a 77% auction information exchange cost for a DSVQC-based actor.

Enabled by the increasingly available embedded hardware accelerators, the capability of executing advanced machine learning models at the edge of the Internet of Things (IoT) triggers interest of applying Artificial Intelligence of Things (AIoT) systems for industrial applications. The in situ inference and decision made based on the sensor data allow the industrial system to address a variety of heterogeneous, local-area non-trivial problems in the last hop of the IoT networks. Such a scheme avoids the wireless bandwidth bottleneck and unreliability issues, as well as the cumbersome cloud. However, the literature still lacks presentations of industrial AIoT system developments that provide insights into the challenges and offer lessons for the relevant research and industry communities. In light of this, we present the design, deployment, and evaluation of an industrial AIoT system for improving the quality control of HP Inc.’s ink cartridge manufacturing lines. While our development has obtained promising results, we also discuss the lessons learned from the whole course of the work, which could be useful to the development of other industrial AIoT systems for quality control in manufacturing.

Data centers (DCs) are power-intensive facilities which use a significant amount of energy for cooling the servers. Increasing the temperature and relative humidity (RH) setpoints is a rule-of-thumb approach to reducing the DC energy usage. However, the high temperature and RH may undermine the server's reliability. Before we can choose the proper temperature and RH settings, it is essential to understand how the temperature and RH setpoints affect the DC power usage and server's reliability. To this end, we constructed and experimented with an air-cooled DC testbed in Singapore, which consists of a direct expansion cooling system and 521 servers running real-world application workloads. This paper presents the key measurement results and observations from our 11-month experiments. Our results suggest that by operating at a supply air temperature setpoints of 29${}^{\circ }$C, our testbed achieves substantial cooling power saving with little impact on the server's reliability. Furthermore, we present a total cost of ownership (TCO) analysis framework which guides settings of the temperature and RH for a DC. Our observations and TCO analysis framework will be useful to future efforts in building and operating air-cooled DCs in tropics and beyond.

Multi-modal fusion has shown initial promising results for object detection of autonomous driving perception. However, many existing fusion schemes do not consider the quality of each fusion input and may suffer from adverse conditions on one or more sensors. While predictive uncertainty has been applied to characterize single-modal object detection performance at run time, incorporating uncertainties into the multi-modal fusion still lacks effective solutions due primarily to the uncertainty’s cross-modal incomparability and distinct sensitivities to various adverse conditions. To fill this gap, this paper proposes Uncertainty-Encoded Mixture-of-Experts (UMoE) that explicitly incorporates single-modal uncertainties into LiDAR-camera fusion. UMoE uses individual expert network to process each sensor’s detection result together with encoded uncertainty. Then, the expert networks’ outputs are analyzed by a gating network to determine the fusion weights. The proposed UMoE module can be integrated into any proposal fusion pipeline. Evaluation shows that UMoE achieves a maximum of 10.67%, 3.17%, and 5.40% performance gain compared with the state-of-the-art proposal-level multi-modal object detectors under extreme weather, adversarial, and blinding attack scenarios.

Tracking interpersonal distances is essential for real-time social distancing management and ex-post contact tracing to prevent spreads of contagious diseases. Bluetooth neighbor discovery has been employed for such purposes in combating COVID-19, but does not provide satisfactory spatiotemporal resolutions. This paper presents ImmTrack, a system that uses a millimeter wave radar and exploits the inertial measurement data from user-carried smartphones or wearables to track interpersonal distances. By matching the movement traces reconstructed from the radar and inertial data, the pseudo identities of the inertial data can be transferred to the radar sensing results in the global coordinate system. The re-identified, radar-sensed movement trajectories are then used to track interpersonal distances. In a broader sense, ImmTrack is the first system that fuses data from millimeter wave radar and inertial measurement units for simultaneous user tracking and re-identification. Evaluation with up to 27 people in various indoor/outdoor environments shows ImmTrack’s decimeters-seconds spatiotemporal accuracy in contact tracing, which is similar to that of the privacy-intrusive camera surveillance and significantly outperforms the Bluetooth neighbor discovery approach.

Existing inertial measurement unit (IMU) based human activity recognition (HAR) approaches still face a major challenge when adopted across users in practice. The severe heterogeneity in IMU data significantly undermines model generalizability in wild adoption. This paper presents UniHAR, a universal HAR framework for mobile devices. To address the challenge of data heterogeneity, we thoroughly study augmenting data with the physics of the IMU sensing process and present a novel adoption of data augmentations for exploiting both unlabeled and labeled data. We consider two application scenarios of UniHAR, which can further integrate federated learning and adversarial training for improved generalization. UniHAR is fully prototyped on the mobile platform and introduces low overhead to mobile devices. Extensive experiments demonstrate its superior performance in adapting HAR models across four open datasets.

Human activity recognition (HAR) will be an essential function of various emerging applications. However, HAR typically encounters challenges related to modality limitations and label scarcity, leading to an application gap between current solutions and real-world requirements. In this work, we propose MESEN, a multimodal-empowered unimodal sensing framework, to utilize unlabeled multimodal data available during the HAR model design phase for unimodal HAR enhancement during the deployment phase. From a study on the impact of supervised multimodal fusion on unimodal feature extraction, MESEN is designed to feature a multi-task mechanism during the multimodal-aided pre-training stage. With the proposed mechanism integrating cross-modal feature contrastive learning and multimodal pseudo-classification aligning, MESEN exploits unlabeled multimodal data to extract effective unimodal features for each modality. Subsequently, MESEN can adapt to downstream unimodal HAR with only a few labeled samples. Extensive experiments on eight public multimodal datasets demonstrate that MESEN achieves significant performance improvements over state-of-the-art baselines in enhancing unimodal HAR by exploiting multimodal data.

In[1], on page 1824,Fig. 3should be as follows:

This paper presents TouchAuth, a new touch-to-access device authentication approach using induced body electric potentials (iBEPs) caused by the indoor ambient electric field that is mainly emitted from the building's electrical network. The design of TouchAuth is based on the electrostatics of iBEP generation and a resulting property, i.e., the iBEPs at two close locations on the same human body are similar, whereas those from different human bodies are distinct. Extensive experiments verify the above property and show that TouchAuth achieves high-profile receiver operating characteristics in implementing the touch-to-access policy. Our experiments also show that a range of possible interfering sources including appliances’ electromagnetic emanations and noise injections into the power network do not affect the performance of TouchAuth. A key advantage of TouchAuth is that the iBEP sensing requires a simple analog-to-digital converter only, which is widely available on microcontrollers. Compared with the existing approaches including intra-body communication and physiological sensing, TouchAuth is a low-cost, faster, and easy-to-use approach for authorized users to access the smart objects found in indoor environments.

Visual sensing has been increasingly employed in various industrial applications including manufacturing process monitoring and worker safety monitoring. This paper presents the design and implementation of a wireless camera system, namely, EFCam, which uses low-power wireless communications and edge-fog computing to achieve cordless and energy-efficient visual sensing. The camera performs image pre-processing and offloads the data to a resourceful fog node for advanced processing using deep models. EFCam admits dynamic configurations of several parameters that form a configuration space. It aims to adapt the configuration to maintain desired visual sensing performance of the deep model at the fog node with minimum energy consumption of the camera in image capture, pre-processing, and data communications, under dynamic variations of the monitored process, the application requirement, and wireless channel conditions. However, the adaptation is challenging due to the complex relationships among the involved factors. To address the complexity, we apply deep reinforcement learning to learn the optimal adaptation policy when a fog node supports one or more wireless cameras. Extensive evaluation based on trace-driven simulations and experiments show that EFCam complies with the accuracy and latency requirements with lower energy consumption for a real industrial product object tracking application, compared with five baseline approaches incorporating hysteresis-based and event-triggered adaptation.

Current LoRa networks including those following the LoRaWAN specification use the primitive ALOHA mechanism for media access control due to LoRa’s lack of carrier sense capability. From our extensive measurements, the channel activity detection feature that was recently introduced to LoRa for energy-efficiently detecting preamble chirps can also detect payload chirps reliably. This sheds light on an efficient carrier-sense multiple access protocol that we refer to as LMAC for LoRa networks. This article presents the designs of three advancing versions of LMAC that respectively implement carrier-sense multiple access, and balance the communication loads among the channels defined by frequencies and spreading factors based on the end nodes’ local information and then additionally the gateway’s global information. Experiments on a 50-node lab testbed and a 16-node university deployment show that, compared with ALOHA, LMAC brings up to 2.2× goodput improvement and 2.4× reduction of radio energy per successfully delivered frame. Thus, should LoRaWAN’s ALOHA be replaced with LMAC, network performance boosts can be realized.

Runtime domain shifts from the training phase caused by sensor characteristic variation incur performance drops of the deep learning-based sensing systems. To address this problem, existing transfer learning techniques require substantial target-domain data and incur high post-deployment overhead. Differently, we propose to exploit the first principle governing the domain shift to reduce the demand for target-domain data. Specifically, our proposed approach called PhyAug uses the first principle fitted with few labeled or unlabeled data pairs collected by the source sensor and the target sensor to transform the existing source-domain training data into the augmented target-domain data for calibrating the deep neural networks. In two audio sensing case studies of keyword spotting and automatic speech recognition, PhyAug recovers the recognition accuracy losses due to microphones’ characteristic variations by 37% to 72% with 5-second unlabeled data collected from the target microphones. In a case study of acoustics-based room recognition, PhyAug recovers the recognition accuracy loss caused by smartphone microphone variation by 33% to 80%. In the last case study of fisheye image recognition, PhyAug reduces the image recognition error due to the camera-induced distortions by 72%.

Indoor localization by leveraging the existing residential instru-ments has been widely explored. Given the properties of tempo-ral stability and spatial distinctness, the electromagnetic radiation (EMR) from the powerline network is a promising signal for location sensing. In this demo, we present a three-sensor setup to capture the powerline EMR signal from the three-dimensional (3D) space and formulate a new powerline EMR feature to implement the simultaneous localization and mapping (SLAM). Compared with the single sensor setup, our proposed approach can improve the localization accuracy to decimeter level.

Massive fraudulent and phishing robocalls present threats to societies. The integration of artificial intelligence technologies, including dialogue and voice generation systems, renders the robocalls more deceptive. Existing countermeasures such as caller ID, call provenance, voiceprint, and fake voice detection have respective limitations and are heavyweight for end users' smartphones. This paper studies detecting the acoustic echo channel on the remote end of a call based on the received voice. The positive detection result evidencing the physical setup of an audio system is indicative of a human caller. However, the acoustic echo cancellation mechanisms of most audio systems and the use of earphone/headset diminish echoes significantly. To address these issues, the proposed Telesonar transmits short chirps during the vulnerable time of echo cancellation, detects the tiny echo remnants from the received voice, and passively analyzes the timing of caller's breath sounds to confirm a human caller. Extensive real experiments under a wide range of settings show that Telesonar correctly recognizes human callers with a rate of over 95%, while wrongly recognizing voice robots as human with a rate of 3.8%.

Mobile cloud offloading is indispensable for inference tasks based on large-scale deep models. However, transmitting privacy-rich inference data to the cloud incurs concerns. This paper presents the design of a system called PriMask, in which the mobile device uses a secret small-scale neural network called MaskNet to mask the data before transmission. PriMask significantly weakens the cloud's capability to recover the data or extract certain private attributes. The MaskNet is cascadable in that the mobile can opt in to or out of its use seamlessly without any modifications to the cloud's inference service. Moreover, the mobiles use different MaskNets, such that the collusion between the cloud and some mobiles does not weaken the protection for other mobiles. We devise a split adversarial learning method to train a neural network that generates a new MaskNet quickly (within two seconds) at run time. We apply PriMask to three mobile sensing applications with diverse modalities and complexities, i.e., human activity recognition, urban environment crowdsensing, and driver behavior recognition. Results show PriMask's effectiveness in all the three applications.

Indoor self-localization is a highly demanded system function for smartphones. The current solutions based on inertial, radio frequency, and geomagnetic sensing may have degraded performance when their limiting factors take effect. In this paper, we present a new indoor simultaneous localization and mapping (SLAM) system that utilizes the smartphone's built-in audio hardware and inertial measurement unit (IMU). Our system uses a smartphone's loud-speaker to emit near-inaudible chirps and then the microphone to record the acoustic echoes from the indoor environment. Our profiling measurements show that the echoes carry location information with sub-meter granularity. To enable SLAM, we apply contrastive learning to construct an echoic location feature (ELF) extractor, such that the loop closures on the smartphone's trajectory can be accurately detected from the associated ELF trace. The detection results effectively regulate the IMU-based trajectory reconstruction. Extensive experiments show that our ELF-based SLAM achieves median localization errors of 0.1 m, 0.53 m, and 0.4m on the reconstructed trajectories in a living room, an office, and a shopping mall, and outperforms the Wi-Fi and geomagnetic SLAM systems.

At present, a co-location data center often applies an identical and low temperature setpoint for its all server rooms. Although increasing the temperature setpoint is a rule-of-thumb approach to reducing the cooling energy usage, the tenants may have different mentalities and technical constraints in accepting higher temperature setpoints. Thus, supporting distinct temperature setpoints is desirable for a co-location data center in pursuing higher energy efficiency. This calls for a new cooling power attribution scheme to address the inter-room heat transfers that can be up to 9% of server load as shown in our real experiments. This article describes our approaches to estimating the inter-room heat transfers, using the estimates to rectify the metered power usages of the rooms’ air handling units, and fairly attributing the power usage of the shared cooling infrastructure (i.e., chiller and cooling tower) to server rooms by following the Shapley value principle. Extensive numeric experiments based on a widely accepted cooling system model are conducted to evaluate the effectiveness of the proposed cooling power attribution scheme. A case study suggests that the proposed scheme incentivizes rational tenants to adopt their highest acceptable temperature setpoints under a non-cooperative game setting. Further analysis considering distinct relative humidity setpoints shows that our proposed scheme also properly and inherently addresses the attribution of humidity control power.

Low-power wide-area network technologies such as long-range wide-area network (LoRaWAN) are promising for collecting low-rate monitoring data from geographically distributed sensors, in which timestamping the sensor data is a critical system function. This article considers a synchronization-free approach to timestamping LoRaWAN uplink data based on signal arrival time at the gateway, which well matches LoRaWAN’s one-hop star topology and releases bandwidth from transmitting timestamps and synchronizing end devices’ clocks at all times. However, we show that this approach is susceptible to a frame delay attack consisting of malicious frame collision and delayed replay. Real experiments show that the attack can affect the end devices in large areas up to about 50,000, m 2 . In a broader sense, the attack threatens any system functions requiring timely deliveries of LoRaWAN frames. To address this threat, we propose a LoRaTS gateway design that integrates a commodity LoRaWAN gateway and a low-power software-defined radio receiver to track the inherent frequency biases of the end devices. Based on an analytic model of LoRa’s chirp spread spectrum modulation, we develop signal processing algorithms to estimate the frequency biases with high accuracy beyond that achieved by LoRa’s default demodulation. The accurate frequency bias tracking capability enables the detection of the attack that introduces additional frequency biases. We also investigate and implement a more crafty attack that uses advanced radio apparatuses to eliminate the frequency biases. To address this crafty attack, we propose a pseudorandom interval hopping scheme to enhance our frequency bias tracking approach. Extensive experiments show the effectiveness of our approach in deployments with real affecting factors such as temperature variations.

Deep learning-based visual sensing has achieved attractive accuracy but is shown vulnerable to adversarial attacks. Specifically, once the attackers obtain the deep model, they can construct adversarial examples to mislead the model to yield wrong classification results. Deployable adversarial examples such as small stickers pasted on the road signs and lanes have been shown effective in misleading advanced driver-assistance systems. Most existing countermeasures against adversarial examples build their security on the attackers’ ignorance of the defense mechanisms. Thus, they fall short of following Kerckhoffs’s principle and can be subverted once the attackers know the details of the defense. This article applies the strategy of moving target defense (MTD) to generate multiple new deep models after system deployment that will collaboratively detect and thwart adversarial examples. Our MTD design is based on the adversarial examples’ minor transferability across different models. The post-deployment of dynamically generated models significantly increase the bar of successful attacks. We also apply serial data fusion with early stopping to reduce the inference time by a factor of up to 5, as well as exploit hardware inference accelerators’ characteristics to strike better tradeoffs between inference time and power consumption. Evaluation based on three datasets including a road sign dataset and two GPU-equipped embedded computing boards shows the effectiveness and efficiency of our approach in counteracting the attack.

Air free cooling is an energy-efficient cooling scheme that has been adopted in the dry and cold climate zones. To adopt this cooling scheme in Singapore's tropical condition, we designed and implemented an air free-cooled DC testbed integrating sensing and control systems for the server and room conditions. Then, we conducted extensive experiments on the testbed to understand its energy efficiency and server reliability. This paper presents the key observations, experiences, and learned lessons obtained from our testbed over a duration of nearly two years. The experiments show that (1) the air free-cooling design can achieve the power usage effectiveness of 1.05, (2) the tropics’ year-round high temperatures up to$37^\circ$C do not impede the air free-cooling, and (3) the implementation of the air free-cooled tropical DCs requires special cares to deal with airborne contaminants to avoid fast corrosion rate and dust-induced server faults. Based on our experiment data, a set of recommendations on the temperature control and the selection of IT equipment for air free-cooled tropical DCs is made. The descriptions of the learned lessons, the resulting recommendations, and the released data can be useful to the relevant research communities, governmental agencies, and standardizing bodies.

Convolutional neural networks (CNNs) are increasingly adopted on resource-constrained sensors for in-situ data analytics in Internet of Things (IoT) applications. This paper presents a model split framework, namely, splitCNN, in order to run a large CNN on a collection of concurrent IoT sensors. Specifically, we adopt CNN filter pruning techniques to split the large CNN into multiple small-size models, each of which is only sensitive to a certain number of data classes. These class-specific models are deployed onto the resource-constrained concurrent sensors which collaboratively perform distributed CNN inference on their same/similar sensing data. The outputs of multiple models are then fused to yield the global inference result. We apply splitCNN to three case studies with different sensing modalities, which include the human voice, industrial vibration signal, and visual sensing data. Extensive evaluation shows the effectiveness of the proposed splitCNN. In particular, the splitCNN achieves significant reduction in the model size and inference time while maintaining similar accuracy, compared with the original CNN model for all three case studies.

Outsourcing the design of deep neural networks may incur cybersecurity threats from the hostile designers. This paper studies a new covert channel attack that leaks the inference results over the air through a hostile design of the neural network architecture and the computing device's electromagnetic radiation when executing the neural network. Specifically, the hostile neural network consists of a series of binary models that correspond to all classes and are executed sequentially. The execution terminates once any binary model given the input is positive about its responsible class. We describe an approach to generate such binary models by pruning a benign neural network that is trained using the standard method to deal with all the classes. Compared with the benign neural network, the hostile one has similar memory usage and negligible classification accuracy drop, but distinct inference times for the samples of different classes. As a result, the hostile neural network's classification result can be eavesdropped by measuring the duration of the electromagnetic radiation emanated from the computing device. As neural networks are stored and transmitted as data files, this covert channel attack is more stealthy to the anti-malware than other code-based attacks. We implement the described attack on two edge computing devices that execute the hostile neural network on CPU or GPU. Evaluation shows 100% empirical accuracy in eavesdropping the inference results.

Run-time domain shifts from training-phase domains are common in sensing systems designed with deep learning. The shifts can be caused by sensor characteristic variations and/or discrepancies between the design-phase model and the actual model of the sensed physical process. To address these issues, existing transfer learning techniques require substantial target-domain data and thus incur high post-deployment overhead. This paper proposes to exploit the first principle governing the domain shift to reduce the demand on target-domain data. Specifically, our proposed approach called PhyAug uses the first principle fitted with few labeled or unlabeled source/target-domain data pairs to transform the existing source-domain training data into augmented data for updating the deep neural networks. In two case studies of keyword spotting and DeepSpeech2-based automatic speech recognition, with 5-second unlabeled data collected from the target microphones, PhyAug recovers the recognition accuracy losses due to microphone characteristic variations by 37% to 72%. In a case study of seismic source localization with TDoA fingerprints, by exploiting the first principle of signal propagation in uneven media, PhyAug only requires 3% to 8% of labeled TDoA measurements required by the vanilla fingerprinting approach in achieving the same localization accuracy.

Enabled by the increasingly available embedded hardware accelerators, the capability of executing advanced machine learning models at the edge of the Internet of Things (IoT) triggers wide interest of applying the resulting Artificial Intelligence of Things (AIoT) systems in industrial applications. The in situ inference and decision made based on the sensor data containing patterns with certain sophistication allow the industrial system to address a variety of heterogeneous, local-area non-trivial problems in the last hop of the IoT networks, avoiding the wireless bandwidth bottleneck and unreliability issues and also the cumbersome cloud. However, the literature still lacks presentations of industrial AIoT system developments that provide insights into the challenges and offer important lessons for the relevant research and engineering communities, no matter the development is successful or not. In light of this, we present the design, deployment, and evaluation of an industrial AIoT system for improving the quality control of Hewlett-Packard's ink cartridge manufacturing lines. While our development has obtained promising results, we also discuss the lessons learned from the whole course of the effort, which could be useful to the developments of other industrial AIoT systems.

Visual sensing has been increasingly employed in industrial processes. This paper presents the design and implementation of an industrial wireless camera system, namely, EFCam, which uses low-power wireless communications and edge-fog computing to achieve cordless and energy-efficient visual sensing. The camera performs image pre-processing (i.e., compression or feature extraction) and transmits the data to a resourceful fog node for advanced processing using deep models. EFCam admits dynamic configurations of several parameters that form a configuration space. It aims to adapt the configuration to maintain desired visual sensing performance of the deep model at the fog node with minimum energy consumption of the camera in image capture, pre-processing, and data communications, under dynamic variations of application requirement and wireless channel conditions. However, the adaptation is challenging due primarily to the complex relationships among the involved factors. To address the complexity, we apply deep reinforcement learning to learn the optimal adaptation policy. Extensive evaluation based on trace-driven simulations and experiments show that EFCam complies with the accuracy and latency requirements with lower energy consumption for a real industrial product object tracking application, compared with four baseline approaches incorporating hysteresis-based adaptation.

Smartphone indoor location awareness is increasingly demanded by a variety of mobile applications. The existing solutions for accurate smartphone indoor localization rely on additional devices or pre-installed infrastructure (e.g., dense WiFi access points, Bluetooth beacons). In this demo, we present EchoLoc, an infrastructure-free smartphone indoor localization system using room acoustic response to a chirp emitted by the phone. EchoLoc consists of a mobile client for echo data collection and a cloud server hosting a deep neural network for location inference. EchoLoc achieves 95% accuracy in recognizing 101 locations in a large public indoor space and a median localization error of 0.5 m in a typical lab area. Demo video is available at https://youtu.be/5si0Cq6LzT4.

Deep learning greatly empowers Inertial Measurement Unit (IMU) sensors for various mobile sensing applications, including human activity recognition, human-computer interaction, localization and tracking, and many more. Most existing works require substantial amounts of well-curated labeled data to train IMU-based sensing models, which incurs high annotation and training costs. Compared with labeled data, unlabeled IMU data are abundant and easily accessible. In this work, we present LIMU-BERT, a novel representation learning model that can make use of unlabeled IMU data and extract generalized rather than task-specific features. LIMU-BERT adopts the principle of self-supervised training of the natural language model BERT to effectively capture temporal relations and feature distributions in IMU sensor measurements. However, the original BERT is not adaptive to mobile IMU data. By meticulously observing the characteristics of IMU sensors, we propose a series of techniques and accordingly adapt LIMU-BERT to IMU sensing tasks. The designed models are lightweight and easily deployable on mobile devices. With the representations learned via LIMU-BERT, task-specific models trained with limited labeled samples can achieve superior performances. We extensively evaluate LIMU-BERT with four open datasets. The results show that the LIMU-BERT enhanced models significantly outperform existing approaches in two typical IMU sensing applications.

Due to the low power of electromagnetic radiation (EMR), EM convert channel has been widely considered as a short-range attack that can be easily mitigated by shielding. This paper overturns this common belief by demonstrating how covert EM signals leaked from typical laptops, desktops and servers are decoded from hundreds of meters away, or penetrate aggressive shield previously considered as sufficient to ensure emission security. We achieve this by designing EMLoRa – a super resilient EM covert channel that exploits memory as a LoRa-like radio. EMLoRa represents the first attempt of designing an EM covert channel using state-of-the-art spread spectrum technology. It tackles a set of unique challenges, such as handling complex spectral characteristics of EMR, tolerating signal distortions caused by CPU contention, and preventing adversarial detectors from demodulating covert signals. Experiment results show that EMLoRa boosts communication range by 20x and improves attenuation resilience by up to 53 dB when compared with prior EM covert channels at the same bit rate. By achieving this, EMLoRa allows an attacker to circumvent security perimeter, breach Faraday cage, and localize air-gapped devices in a wide area using just a small number of inexpensive sensors. To countermeasure EMLoRa, we further explore the feasibility of uncovering EMLoRa's signal using energy- and CNN-based detectors. Experiments show that both detectors suffer limited range, allowing EMLoRa to gain a significant range advantage. Our results call for further research on the countermeasure against spread spectrum-based EM covert channels.

Home appliance manufacturers strive to obtain feedback from users to improve their products and services to build a smart home system. To help manufacturers develop a smart home system, we design a federated learning (FL) system leveraging a reputation mechanism to assist home appliance manufacturers to train a machine learning model based on customers’ data. Then, manufacturers can predict customers’ requirements and consumption behaviors in the future. The working flow of the system includes two stages: in the first stage, customers train the initial model provided by the manufacturer using both the mobile phone and the mobile-edge computing (MEC) server. Customers collect data from various home appliances using phones, and then they download and train the initial model with their local data. After deriving local models, customers sign on their models and send them to the blockchain. In case customers or manufacturers are malicious, we use the blockchain to replace the centralized aggregator in the traditional FL system. Since records on the blockchain are untampered, malicious customers or manufacturers’ activities are traceable. In the second stage, manufacturers select customers or organizations as miners for calculating the averaged model using received models from customers. By the end of the crowdsourcing task, one of the miners, who is selected as the temporary leader, uploads the model to the blockchain. To protect customers’ privacy and improve the test accuracy, we enforce differential privacy (DP) on the extracted features and propose a new normalization technique. We experimentally demonstrate that our normalization technique outperforms batch normalization when features are under DP protection. In addition, to attract more customers to participate in the crowdsourcing FL task, we design an incentive mechanism to award participants.

The Internet of Things (IoT) will be a main data generation infrastructure for achieving better system intelligence. This article considers the design and implementation of a practical privacy-preserving collaborative learning scheme, in which a curious learning coordinator trains a better machine learning model based on the data samples contributed by a number of IoT objects, while the confidentiality of the raw forms of the training data is protected against the coordinator. Existing distributed machine learning and data encryption approaches incur significant computation and communication overhead, rendering them ill-suited for resource-constrained IoT objects. We study an approach that applies independent random projection at each IoT object to obfuscate data and trains a deep neural network at the coordinator based on the projected data from the IoT objects. This approach introduces light computation overhead to the IoT objects and moves most workload to the coordinator that can have sufficient computing resources. Although the independent projections performed by the IoT objects address the potential collusion between the curious coordinator and some compromised IoT objects, they significantly increase the complexity of the projected data. In this article, we leverage the superior learning capability of deep learning in capturing sophisticated patterns to maintain good learning performance. Extensive comparative evaluation shows that this approach outperforms other lightweight approaches that apply additive noisification for differential privacy and/or support vector machines for learning in the applications with light to moderate data pattern complexities.

Air free-cooled data centers (DCs) have not existed in the tropical zone due to the unique challenges of year-round high ambient temperature and relative humidity (RH). The increasing availability of servers that can tolerate higher temperatures and RH due to the regulatory bodies’ prompts to raise DC temperature setpoints sheds light upon the feasibility of air free-cooled DCs in the tropics. However, due to the complex psychrometric dynamics, operating the air free-cooled DC in the tropics generally requires adaptive control of supply air condition to maintain the computing performance and reliability of the servers. This article studies the problem of controlling the supply air temperature and RH in a free-cooled tropical DC below certain thresholds. To achieve the goal, we formulate the control problem as Markov decision processes and apply deep reinforcement learning (DRL) to learn the control policy that minimizes the cooling energy while satisfying the requirements on the supply air temperature and RH. We also develop a constrained DRL solution for performance improvements. Extensive evaluation based on real data traces collected from an air free-cooled testbed and comparisons among the unconstrained and constrained DRL approaches as well as two other baseline approaches show the superior performance of our proposed solutions.

Low-power wide-area network technologies such as LoRaWAN are promising for collecting low-rate monitoring data from geographically distributed sensors, in which timestamping the sensor data is a critical system function. This paper considers a synchronization-free approach to timestamping LoRaWAN uplink data based on signal arrival time at the gateway, which well matches LoRaWAN’s one-hop star topology and releases bandwidth from transmitting timestamps and synchronizing end devices’ clocks at all times. However, we show that this approach is susceptible to a frame delay attack consisting of malicious frame collision and delayed replay. Real experiments show that the attack can affect the end devices in large areas up to about 50, 000 m2. In a broader sense, the attack threatens any system functions requiring timely deliveries of LoRaWAN frames. To address this threat, we propose a LoRaTS gateway design that integrates a commodity LoRaWAN gateway and a low-power software-defined radio receiver to track the inherent frequency biases of the end devices. Based on an analytic model of LoRa’s chirp spread spectrum modulation, we develop signal processing algorithms to estimate the frequency biases with high accuracy beyond that achieved by LoRa’s default demodulation. The accurate frequency bias tracking capability enables the detection of the attack that introduces additional frequency biases. Extensive experiments show the effectiveness of our approach.

Current LoRa networks including those following the LoRaWAN specification use the primitive ALOHA mechanism for media access control due to LoRa's lack of carrier sense capability. From our extensive measurements, the Channel Activity Detection (CAD) feature that is recently introduced to LoRa for energy-efficiently detecting preamble chirps, can also detect payload chirps reliably. This sheds light on an efficient carrier-sense multiple access (CSMA) protocol that we call LMAC for LoRa networks. This paper presents the designs of three advancing versions of LMAC that respectively implements CSMA, balances the communication loads among the channels defined by frequencies and spreading factors based on the end nodes' local information and then additionally the gateway's global information. Experiments on a 50-node lab testbed and a 16-node university deployment show that, compared with ALOHA, LMAC brings up to 2.2× goodput improvement and 2.4× reduction of radio energy per successfully delivered frame. Thus, should the LoRaWAN's ALOHA be replaced with LMAC, network performance boosts can be realized.

While thriving application (app) distribution systems involving incentivized third-party app vendors are desirable for the emerging edge computing paradigm, they also bring security challenges as faced by the current mobile app distribution systems. This article studies a threat calledcovert device association, in which the vendors of two apps collude to figure out which of their app installations run on the same edge device. The threat can widely spread when the two apps are popular. It is also a stepping stone for: 1) the de-anonymization attacks against the users anonymous to one of the two vendors and 2) privilege escalation in which the two colluding vendors have united privileges. We show that the threat can be implemented via a reliable and ubiquitous covert channel based on the edge device’s processor workload without requiring any privileged permissions. We present the implementation details for three attack scenarios of: 1) two Android apps; 2) an Android app and a Web session running in the mobile Tor browser; and 3) two Android Things apps. Evaluation on two smartphones and an embedded edge device shows that the covert channel gives at least 0.25 b/s data rate with zero empirical bit error rate and the covert device association can be completed within 3.2 min.

Executing deep neural networks for inference on the server-class or cloud backend based on the data generated at the edge of the Internet of Things is desirable due primarily to the limited compute power of the edge devices and the need to protect the confidentiality of the inference neural networks. However, such a remote inference scheme incurs concerns regarding the privacy of the inference data transmitted by the edge devices to the curious backend. This article presents a lightweight and unobtrusive approach to obfuscate the inference data at the edge devices. It is lightweight in that the edge device only needs to execute a small-scale neural network; it is unobtrusive in that the edge device does not need to indicate whether obfuscation is applied. Extensive evaluation by three case studies of free-spoken digit recognition, handwritten digit recognition, and American sign language recognition shows that our approach effectively protects the confidentiality of the raw forms of the inference data while effectively preserving backend's inference accuracy.

Due to recent cyber attacks on various cyber-physical systems (CPSes), traditional isolation based security schemes in the critical systems are insufficient to deal with the smart adversaries in CPSes with advanced information and communication technologies (ICTs). In this paper, we develop real-time assessment and mitigation of an attack's impact as a system's built-in mechanisms. We study a general class of attacks, which we call time delay attack, that delays the transmissions of control data packets in the CPS control loops. Based on a joint stability-safety criterion, we propose the attack impact assessment consisting of (i) a machine learning (ML) based safety classification, and (ii) a tandem stability-safety classification that exploits a basic relationship between stability and safety, namely that an unstable system must be unsafe whereas a stable system may not be safe. In this assessment approach, the ML addresses a state explosion problem in the safety classification, whereas the tandem structure reduces false negatives in detecting unsafety arising from imperfect ML. We apply our approach to assess the impact of the attack on power grid automatic generation control, and accordingly develop a two-tiered mitigation that tunes the control gain automatically to restore safety where necessary and shed load only if the tuning is insufficient. We also apply our attack impact assessment approach to a thermal power plant control system consisting of two PID control loops. A mitigation approach by tuning the PID controller is also proposed. Extensive simulations based on a 37-bus system model and a thermal power plant control system are conducted to evaluate the effectiveness of our assessment and mitigation approaches.

Naturally occurring disturbances and malicious attacks can lead to faults in synchronizing the clocks of two network nodes. In this article, we investigate the fundamental resilience bounds of network clock synchronization for a system of N nodes against the peer-to-peer synchronization faults. Our analysis is based on practical synchronization algorithms with time complexity down to O ( N 3 ) that attempt to correct the faults by checking the consistency among the following three types of data: (1) the estimated faults, (2) the estimated clock offsets among the nodes, and (3) the measured clock offsets from the potentially faulty peer-to-peer synchronization sessions. Our analysis gives the following three major results. First, the maximum number of faults that can be corrected by the algorithms has a tight bound of ⌊ N /2 ⌋ − 1 when every node pair performs a synchronization session. Second, by converting the fault resilience problem to a graph-theoretic edge connectivity problem and applying Menger’s theorem, we develop an algorithm to compute the tight bound when not every node pair performs a synchronization session. Third, the number of synchronization sessions to achieve the capability of correcting any K faults has a lower bound of ⌈ N (2 K +1) / 2 ⌉ ; we also develop an algorithm to schedule the synchronization sessions to approach the lower bound. The above results provide basic understanding and useful guidelines to the design of resilient clock synchronization systems. For instance, our results suggest that, the four-node network achieves the highest degree of resilience that is defined as the ratio of the maximum number of correctable faults to the number of synchronization sessions. Therefore, by organizing a large-scale clock synchronization system into a hierarchy of multiple tiers with each consisting of four-node synchronization groups, we can achieve satisfactory and understood resilience against faults with reduced synchronization sessions.

Reducing the energy consumption of the servers in a data center via proper job allocation is desirable. Existing advanced job allocation algorithms, based on constrained optimization formulations capturing servers' complex power consumption and thermal dynamics, often scale poorly with the data center size and optimization horizon. This article applies deep reinforcement learning to build an allocation algorithm for long-lasting and compute-intensive jobs that are increasingly seen among today's computation demands. Specifically, a deep Q-network is trained to allocate jobs, aiming to maximize a cumulative reward over long horizons. The training is performed offline using a computational model based on long short-term memory networks that capture the servers' power and thermal dynamics. This offline training approach avoids slow online convergence, low energy efficiency, and potential server overheating during the agent's extensive state-action space exploration if it directly interacts with the physical data center in the usually adopted online learning scheme. At run time, the trained Q-network is forward-propagated with little computation to allocate jobs. Evaluation based on eight months' physical state and job arrival records from a national supercomputing data center hosting 1,152 processors shows that our solution reduces computing power consumption by more than 10 percent and processor temperature by more than 4°C without sacrificing job processing throughput.

Reducing the energy consumption of the servers in a data center via proper job allocation is desirable. Existing advanced job allocation algorithms, based on constrained optimization formulations capturing servers' complex power consumption and thermal dynamics, often scale poorly with the data center size and optimization horizon. This paper applies deep reinforcement learning (DRL) to build an allocation algorithm for long-lasting and compute-intensive jobs that are increasingly seen among today's computation demands. Specifically, a deep Q-network is trained to allocate jobs, aiming to maximize a cumulative reward over long horizons. The training is performed offline using a computational model based on long short-term memory networks that capture the servers' power and thermal dynamics. This offline training approach avoids slow online convergence, low energy efficiency, and potential server overheating during the DRL's extensive state-action space exploration if it directly interacts with the physical data center in the usually adopted online learning scheme. At run time, the trained Q-network is forward-propagated with little computation to allocate jobs. Evaluation based on 8 months' physical state and job arrival records from a national supercomputing data center hosting 1,152 processors shows that our solution reduces computing power consumption by nearly 10% and processor temperature by more than 3°C without sacrificing job processing throughput.

LoRa is an emerging low-power wide-area network technology. Existing studies have focused on LoRa's communication performance. Differently, we study two physical properties of LoRa, i.e., its performance in timing the signal propagation and the transmitters' frequency traits. Signal timing is a basis for implementing clock synchronization, ranging, and advanced physical (PHY) layer techniques such as concurrent decoding. However, LoRa end devices do not provide PHY-layer timestamping that is needed for accurate timing. We propose a SoftLoRa design that integrates a low-power software-defined radio receiver with a LoRa transceiver to provide PHY-layer access. Experiments show that SoftLoRa achieves microseconds timing accuracy over one kilometer and in a multistory building with strong signal attenuation.

This paper presents TouchAuth, a new touch-to-access device authentication approach using induced body electric potentials (iBEPs) caused by the indoor ambient electric field that is mainly emitted from the building's electrical cabling. The design of TouchAuth is based on the electrostatics of iBEP generation and a resulting property, i.e., the iBEPs at two close locations on the same human body are similar, whereas those from different human bodies are distinct. Extensive experiments verify the above property and show that TouchAuth achieves high-profile receiver operating characteristics in implementing the touch-to-access policy. Our experiments also show that a range of possible interfering sources including appliances' electromagnetic emanations and noise injections into the power network do not affect the performance of TouchAuth. A key advantage of TouchAuth is that the iBEP sensing requires a simple analog-to-digital converter only, which is widely available on microcontrollers. Compared with existing approaches including intra-body communication and physiological sensing, TouchAuth is a low-cost, lightweight, and convenient approach for authorized users to access the smart objects found in indoor environments.

Recent technological advances have fostered the development of complex industrial cyber-physical systems with communication delay requirements. The consequences of delay requirement violation in such systems may become increasingly severe. In this paper, we propose a contract-based fault-resilient methodology which aims at managing the communication delays of network flows in industries. With this objective, we present a lightweight mechanism to estimate end-to-end communication delays in the network where the clocks of the switches are not synchronized. The mechanism aims at providing high level of accuracy with little communication overhead. We then propose a contract-based framework using software-defined networking (SDN) where the components are associated with delay contracts and a resilience manager. The proposed resilience management framework contains: (1) contracts which state requirements about components' behaviors, (2) observers which are responsible to detect contract failure (fault), (3) monitors to detect events such as run-time changes in the delay requirements and link failure, (4) control logic to take suitable decisions based on the type of the fault, (5) resilience manager to decide response strategies containing the best course of action as per the control logic decision. Finally, we present a delay-aware path finding algorithm which is used to route/reroute the network flows to provide resilience in the case of faults and, to adapt to the changes in the network state. Performance of the proposed framework is evaluated with the Ryu SDN controller and Mininet network emulator.

Deep learning-based visual sensing has achieved attractive accuracy but is shown vulnerable to adversarial example attacks. Specifically, once the attackers obtain the deep model, they can construct adversarial examples to mislead the model to yield wrong classification results. Deployable adversarial examples such as small stickers pasted on the road signs and lanes have been shown effective in misleading advanced driver-assistance systems. Many existing countermeasures against adversarial examples build their security on the attackers' ignorance of the defense mechanisms. Thus, they fall short of following Kerckhoffs's principle and can be subverted once the attackers know the details of the defense. This paper applies the strategy of moving target defense (MTD) to generate multiple new deep models after system deployment, that will collaboratively detect and thwart adversarial examples. Our MTD design is based on the adversarial examples' minor transferability across different models. The post-deployment dynamically generated models significantly increase the bar of successful attacks. We also apply serial data fusion with early stopping to reduce the inference time by a factor of up to 5. Evaluation based on four datasets including a road sign dataset and two GPU-equipped Jetson embedded computing platforms shows the effectiveness of our approach.

Many clock synchronization protocols based on message passing, e.g., the Network Time Protocol (NTP), assume symmetric network delays to estimate the one-way packet transmission time as half of the round-trip time. As a result, asymmetric network delays caused by either network congestion or malicious packet delays can cause significant synchronization errors. This article exploits sinusoidal voltage signals of an alternating current (AC) power grid to limit the impact of the asymmetric network delays on these clock synchronization protocols. Our extensive measurements show that the voltage signals at geographically distributed locations in a city are highly synchronized. Leveraging calibrated voltage phases, we develop a new clock synchronization protocol that we call Grid Time Protocol (GTP), which allows direct measurement of one-way packet transmission times between its slave and master nodes, subject to an analytic condition that can be easily verified in practice. The direct measurements render GTP resilient against asymmetric network delays under this condition. A prototype implementation of GTP maintains sub-millisecond synchronization accuracy for two nodes tens of kilometers apart in the presence of malicious packet delays. The result has been demonstrated for both Singapore and Hangzhou, China. Simulations driven by real network delay measurements between Singapore and Hangzhou under both normal and congested network conditions also show the synchronization accuracy improvement by GTP. We believe that GTP is suitable for grid-connected distributed systems that are currently served by NTP but desire higher resilience against unfavorable network dynamics and packet delay attacks.

Delivering High-Definition (HD) real-time video to mobile devices is challenged with stringent constraints in delay and reliability. In the presence of network dynamics (e.g., channel errors and bandwidth fluctuations), the existing error-control mechanisms [e.g., Automatic Repeat reQuest (ARQ) and Forward Error Correction (FEC)] frequently induce deadline violations and quality degradations. To strike an effective balance between delay and reliability in real-time video transmission, this research presents an application-layer solution dubbed Partial rEliability based Real-timEStreaming (PERES) to perform partially reliable transfer. First, we develop an analytical framework to model the delay-constrained partial reliability for acknowledgement (ACK) and negative acknowledgement (NAK) based real-time video streaming. Second, we propose scheduling algorithms for video-aware reliability adaptation and network-adaptive buffer control. PERES is able to maximize the transmission reliability of high-priority video frames within stringent delay constraint. We implement the proposed transmission scheme in embedded video monitoring systems and evaluate the efficacy over different wireless network environments. Evaluation results demonstrate PERES achieves appreciable improvements over the reference schemes in perceptual video quality, delay performance, and bandwidth efficiency.

Design of clock synchronization for networked nodes faces a fundamental trade-off between synchronization accuracy and universality of heterogeneous platforms, because a high synchronization accuracy generally requires platform-dependent hardware-level network packet timestamping. This paper presents TouchSync, a new indoor clock synchronization approach for wearables that achieves millisecond accuracy while preserving universality in that it uses standard system calls only, such as reading system clock, sampling sensors, and sending/receiving network messages. The design of TouchSync is driven by a key finding from our extensive measurements that the skin electric potentials (SEPs) induced by powerline radiation are salient, periodic, and synchronous on a same wearer and even across different wearers. TouchSync integrates the SEP signal into the universal principle of Network Time Protocol and solves an integer ambiguity problem by fusing the ambiguous results in multiple synchronization rounds to conclude an accurate clock offset between two synchronizing wearables. With our shared code, TouchSync can be readily integrated into any wearable applications. Extensive evaluation based on our Arduino and TinyOS implementations shows that TouchSync's synchronization errors are below 3 and 7 milliseconds on the same wearer and between two wearers 10 kilometers apart, respectively.

Prompted by technological advancements in wireless systems and handheld devices, concurrent multipath transfer is a promising solution to stream high-quality mobile videos in heterogeneous access medium. Multipath TCP (MPTCP) is a transport-layer protocol recommended by the Internet Engineering Task Force (IETF) for concurrent data transmission to multi-radio terminals. However, it is still challenging to stream high-quality real-time videos with the existing MPTCP solutions because of the tradeoff between energy efficiency and video quality. To deliver real-time video in an energy-efficient manner, this paper presents a Delay-Energy-quAlity-aware MPTCP (DEAM) solution. First, an analytical framework is developed to characterize the delay-constrained energy-quality tradeoff for multipath video delivery over heterogeneous access networks. Second, a subflow allocation algorithm is proposed to minimize the device energy consumption while achieving target video quality within the imposed deadline. The performance of the proposed DEAM is verified by means of extensive Exata emulations with real-time streaming videos. Evaluation results demonstrate that DEAM achieves appreciable improvements over the reference MPTCP solutions in mobile energy conservation and user-perceived video quality.

Separation of Control and Data Planes (SCDP) is a desirable paradigm for low-power multi-hop wireless sensor networks requiring high network performance and manageability. Existing SCDP networks generally adopt an in-band control plane scheme in that the control-plane messages are delivered by their data-plane networks. The physical coupling of the two planes may lead to undesirable consequences. Recently, multi-radio platforms (e.g., TI CC1350 and OpenMote B) are increasingly available, which make the physical separation of the control and data planes possible. To advance the network architecture design, we propose to leverage on the long-range communication capability of the Low-Power Wide-Area Network (LPWAN) radios to form one-hop out-of-band control planes. LoRaWAN, an open, inexpensive, and ISM band based LPWAN radio, is chosen to prototype our out-of-band control plane called LoRaCP. Several characteristics of LoRaWAN such as downlink-uplink asymmetry and primitive ALOHA media access control need to be dealt with to achieve high reliability and efficiency. To address these challenges, a TDMA-based multi-channel transmission control is designed, which features an urgent channel and negative acknowledgment. On a testbed of 16 nodes, LoRaCP is applied to physically separate the control-plane network of the Collection Tree Protocol (CTP) from its Zigbee-based data-plane network. Extensive experiments show that LoRaCP increases CTP’s packet delivery ratio from 65% to 80% in the presence of external interference, while consuming a per-node average radio power of 2.97mW only.

Recent studies exploited external periodic synchronous signals to synchronize a pair of network nodes to address a threat of delaying the communications between the nodes. However, the sensing-based synchronization may yield faults due to nonmalicious signal and sensor noises. This paper considers a system of N nodes that will fuse their peer-to-peer synchronization results to correct the faults. Our analysis gives the lower bound of the number of faults that the system can tolerate when N is up to 12. If the number of faults is no greater than the lower bound, the faults can be identified and corrected. We also prove that the system cannot tolerate more than N - 2 faults. Our results can guide the design of resilient sensing-based clock synchronization systems.

Separation of control and data planes (SCDP) is a desirable paradigm for low-power multi-hop wireless networks requiring high network performance and manageability. Existing SCDP networks generally adopt an in-band control plane scheme in that the control-plane messages are delivered by their data-plane networks. The physical coupling of the two planes may lead to undesirable consequences. To advance the network architecture design, we propose to leverage on the long-range communication capability of the increasingly available low-power wide-area network (LPWAN) radios to form one-hop out-of-band control planes. We choose LoRaWAN, an open, inexpensive, and ISM band based LPWAN radio to prototype our out-of-band control plane called LoRaCP. Several characteristics of LoRaWAN such as downlink-uplink asymmetry and primitive ALOHA media access control (MAC) present challenges to achieving reliability and efficiency. To address these challenges, we design a TDMA-based multi-channel MAC featuring an urgent channel and negative acknowledgment. On a testbed of 16 nodes, we demonstrate applying LoRaCP to physically separate the control-plane network of the Collection Tree Protocol (CTP) from its ZigBee-based data-plane network. Extensive experiments show that LoRaCP increases CTP's packet delivery ratio from 65 % to 80 % in the presence of external interference, while consuming a per-node average radio power of 2.97mW only.

Various sensing modalities have been exploited for indoor location sensing, each of which has well understood limitations, however. This paper presents a first systematic study on using the electromagnetic field (EMF) induced by a building's electric power network for simultaneous localization and mapping (SLAM). A basis of this work is a measurement study showing that the power network EMF sensed by either a customized sensor or smartphone's microphone as a side-channel sensor is spatially distinct and temporally stable. Based on this, we design a SLAM approach that can reliably detect loop closures based on EMF sensing results. With the EMF feature map constructed by SLAM, we also design an efficient online localization scheme for resource-constrained mobiles. Evaluation in three indoor spaces shows that the power network EMF is a promising modality for location sensing on mobile devices, which is able to run in real time and achieve sub-meter accuracy.

The rapid proliferation of wireless video cameras has raised serious privacy concerns. In this paper, we propose a stimulating-and-probing approach to detecting wireless spy cameras. The core idea is to actively alter the light condition of a private space to manipulate the spy camera's video scene, and then investigates the responsive variations of a packet flow to determine if it is produced by a wireless camera. Following this approach, we develop Blink and Flicker -- two practical systems for detecting wireless spy cameras. Blink is a lightweight app that can be deployed on off-the-shelf mobile devices. It asks the user to turn on/off the light of her private space, and then uses the light sensor and the wireless radio of the mobile device to identify the response of wireless cameras. Flicker is a robust and automated system that augments Blink to detect wireless cameras in both live and offline streaming modes. Flicker employs a cheap and portable circuit, which harnesses daily used LEDs to stimulate wireless cameras using human-invisible flickering. The time series of stimuli is further encoded using FEC to combat ambient light and uncontrollable packet flow variations that may degrade detection performance. Extensive experiments show that Blink and Flicker can accurately detect wireless cameras under a wide range of network and environmental conditions.

Modern urban railways extensively use computerized sensing and control technologies to achieve safe, reliable, and well-timed operations. However, the use of these technologies may provide a convenient leverage to cyber-attackers who have bypassed the air gaps and aim at causing safety incidents and service disruptions. In this article, we study False Data Injection (FDI) attacks against railway Traction Power Systems (TPSes). Specifically, we analyze two types of FDI attacks on the train-borne voltage, current, and position sensor measurements—which we call efficiency attack and safety attack— that (i) maximize the system’s total power consumption and (ii) mislead trains’ local voltages to exceed given safety-critical thresholds, respectively. To counteract, we develop a Global Attack Detection (GAD) system that serializes a bad data detector and a novel secondary attack detector designed based on unique TPS characteristics. With intact position data of trains, our detection system can effectively detect FDI attacks on trains’ voltage and current measurements even if the attacker has full and accurate knowledge of the TPS, attack detection, and real-time system state. In particular, the GAD system features an adaptive mechanism that ensures low false-positive and negative rates in detecting the attacks under noisy system measurements. Extensive simulations driven by realistic running profiles of trains verify that a TPS setup is vulnerable to FDI attacks, but these attacks can be detected effectively by the proposed GAD while ensuring a low false-positive rate.

The continuous fluctuation of electric network frequency (ENF) presents a fingerprint indicative of time, which we call natural timestamp . This article studies the time accuracy of these natural timestamps obtained from powerline electromagnetic radiation (EMR), which is mainly excited by powerline voltage oscillations at the rate of the ENF. However, since the EMR signal is often weak and noisy, extracting the ENF is challenging, especially on resource-limited sensor platforms. We design an efficient EMR conditioning algorithm and evaluate the time accuracy of EMR natural timestamps on two representative classes of IoT platforms—a high-end single-board computer with a customized EMR antenna and a low-end mote with a normal conductor wire acting as EMR antenna. Extensive measurements at six sites in a city, which are away from each other for up to 24km, show that the high-end and low-end nodes achieve median time errors of about 50ms and 150ms, respectively. To demonstrate the use of the EMR natural timestamps, we discuss three applications: time recovery, runtime clock verification, and secure clock synchronization.

Desynchronized clocks among network nodes in critical infrastructures can degrade system performance and even lead to safety incidents. Clock synchronization protocols based on network message exchanges, though widely used in current network systems, are susceptible to delay attacks against the packet transmission. This vulnerability cannot be solved by conventional security measures, such as encryption, and remains an open problem. This article proposes to use the sine voltage waveform of a utility power grid to synchronize network nodes connected to the same grid. Our experiments demonstrate that minute fluctuations of the voltage’s cycle length encode fine-grained global time information in Singapore’s utility grid. Based on this key result, we develop a clock synchronization approach that achieves good accuracy and is provably secure against packet-delay attacks. Implementation results show that our approach achieves an average synchronization error of 0.1 ms between two network nodes that are deployed in office and residential buildings 10 km apart. When the proposed system is deployed within the same floor of an office building, the error reduces to 10 μs. When there are heavy industrial loads close to one of the two nodes 10 km apart, the system can still maintain subsecond accuracy. Moreover, when the two nodes are deployed within the same building floor with industrial loads nearby, the average synchronization error is 34 μ

Many clock synchronization protocols based on message passing, e.g., the Network Time Protocol (NTP), assume symmetric network delays to estimate the one-way packet transmission time as half of the round-trip time. As a result, asymmetric network delays caused by either %natural one-way network congestion or malicious packet delays can cause significant synchronization errors. This paper exploits sinusoidal voltage signals of an alternating current (ac) power grid to tame the asymmetric network delays for robust and resilient clock synchronization. Our extensive measurements show that the voltage signals at geographically distributed locations in a city are highly synchronized. Leveraging calibrated voltage phases, we develop a new clock synchronization protocol, which we call Grid Time Protocol (GTP), that allows direct measurement of one-way packet transmission times between its slave and master nodes, under an analytic condition that can be easily verified in practice. The direct measurements render GTP resilient against asymmetric network delays under this condition. A prototype implementation of GTP, based on readily available ac/ac transformers and PC-grade sound cards as voltage signal sampling devices, maintains sub-ms synchronization accuracy for two nodes 30 km apart, in the presence of malicious packet delays. We believe that GTP is suitable for grid-connected distributed systems that are currently served by NTP but desire higher resilience against network dynamics and packet delay attacks.

Increasing dynamics of electrical loads presents uncertainty and hence new challenges for power grid controls and optimization. In economic dispatch control (EDC) for minimizing generation cost, demand reporting by customers is a promising approach for managing the uncertainty, but it raises important privacy concerns. Adding random noise to aggregate queries of demand reports can provide differential privacy (DP) for the individual customers. But the noisy query results can adversely impact the EDC's optimality. In this paper, we analyze the privacy cost in demand reporting in terms of how DP-induced noise will increase the total generation cost. Our analysis shows that the noise amounts for different customers are intricately coupled with one another in determining the total cost. In view of the coupling, we apply the principle of Shapley value to attribute fair shares of the total cost to the power grid buses. For efficient sharing of the privacy cost, in a manner scalable to large power systems with many buses, we additionally propose heuristic algorithms to approximate the Shapley value. Trace-driven simulations based on a 5-bus power system model validate our analysis and illustrate the performance of the proposed cost sharing algorithms.

The continuous fluctuation of electric network frequency (ENF) presents a fingerprint indicative of time, which we call natural timestamp. This paper studies the time accuracy of these natural timestamps obtained from powerline electromagnetic radiation (EMR), which is mainly excited by powerline voltage oscillations at the rate of the ENF. However, since the EMR signal is often weak and noisy, extracting the ENF is challenging, especially on resource-limited sensor platforms. We design an efficient EMR conditioning algorithm and evaluate the time accuracy of EMR natural timestamps on two representative classes of IoT platforms - a high-end single-board computer with a customized EMR antenna and a low-end mote with a normal conductor wire acting as EMR antenna. Extensive measurements at five sites in a city, which are away from each other for up to 24 km, show that the high-end and low-end nodes achieve median time errors of about 50 ms and 150 ms, respectively. To demonstrate the use of the EMR natural timestamps, we discuss two applications, namely time recovery and runtime clock verification.

The continuous fluctuation of electric network frequency (ENF) presents a fingerprint indicative of time, which we call natural timestamp. This live demo demonstrates the accuracy of the natural timestamps obtained by four wired voltage sensors and four wireless electromagnetic radiation (EMR) sensors that are geographically distributed in Singapore. The voltage sensors and the EMR sensors capture the minute fluctuations of the length of each voltage cycle and the average ENF over every 50 voltage cycles, respectively. The evaluation in our prior studies [1, 3] has shown that the natural timestamps recorded by the voltage sensors and the EMR sensors give sub-millisecond and sub-second average time errors, respectively. This demo will also show their time errors.

Design of clock synchronization for networked nodes faces a fundamental trade-off between synchronization accuracy and universality for heterogeneous platforms, because a high synchronization accuracy generally requires platform-dependent hardware-level network packet timestamping. This paper presents TouchSync, a new indoor clock synchronization approach for wearables that achieves millisecond accuracy while preserving universality in that it uses standard system calls only, such as reading system clock, sampling sensors, and sending/receiving network messages. The design of TouchSync is driven by a key finding from our extensive measurements that the skin electric potentials (SEPs) induced by powerline radiation are salient, periodic, and synchronous on a same wearer and even across different wearers. TouchSync integrates the SEP signal into the universal principle of Network Time Protocol and solves an integer ambiguity problem by fusing the ambiguous results in multiple synchronization rounds to conclude an accurate clock offset between two synchronizing wearables. With our shared code, TouchSync can be readily integrated into any wearable applications. Extensive evaluation based on our Arduino and TinyOS implementations shows that TouchSync's synchronization errors are below 3 and 7 milliseconds on the same wearer and between two wearers 10 kilometers apart, respectively.

Load shedding can combat the overload of a power grid that may jeopardize the grid’s safety. However, disconnected customers may be excessively inconvenienced or even endangered. With the emergence of demand-response based on cyber-enabled smart meters and appliances, customers may participate in solving the overload by curtailing their demands collaboratively such that no single customers will have to bear a disproportionate burden of reduced usage. However, compliance or commitment to curtailment requests by untrusted users is uncertain, which causes an important safety concern. This article proposes a two-phase load management scheme that (i) gives customers a chance to curtail their demands and correct a grid’s overload when there are no immediate safety concerns but (ii) falls back to load shedding to ensure safety once the grid enters a vulnerable state. Extensive simulations based on a 37-bus electrical grid and traces of real electrical load demonstrate the effectiveness of this scheme. In particular, if customers are, as expected, sufficiently committed to the load curtailment, overloads can be resolved in real time by collaborative and graceful usage degradation among them, thereby avoiding unpleasant load shedding.

This paper studies the impact of false data injection (FDI) attacks on automatic generation control (AGC), a fundamental control system used in all power grids to maintain the grid frequency at a nominal value. Attacks on the sensor measurements for AGC can cause frequency excursion that triggers remedial actions, such as disconnecting customer loads or generators, leading to blackouts, and potentially costly equipment damage. We derive an attack impact model and analyze an optimal attack, consisting of a series of FDIs that minimizes the remaining time until the onset of disruptive remedial actions, leaving the shortest time for the grid to counteract. We show that, based on eavesdropped sensor data and a few feasible-to-obtain system constants, the attacker can learn the attack impact model and achieve the optimal attack in practice. This paper provides essential understanding on the limits of physical impact of the FDIs on power grids, and provides an analysis framework to guide the protection of sensor data links. For countermeasures, we develop efficient algorithms to detect the attack, estimate which sensor data links are under attack, and mitigate attack impact. Our analysis and algorithms are validated by experiments on a physical 16-bus power system test bed and extensive simulations based on a 37-bus power system model.

Owing to the rich processing, multi-modal sensing, and versatile networking capabilities, smartphones are increasingly used to build data-intensive embedded sensing applications. However, various challenges must be systematically addressed before smartphones can be used as a generic embedded sensing platform, including high power consumption, lack of real-time functionality, and user-friendly embedded programming support. This paper presents ORBIT, a smartphone-based platform for data-intensive embedded sensing applications. ORBIT features a tiered architecture, in which a smartphone can interface to an energy-efficient peripheral board and/or a cloud service. ORBIT as a platform addresses the shortcomings of current smartphones while utilizing their strengths. ORBIT provides a profile-based task partitioning that allows it to intelligently dispatch the processing tasks among the tiers to minimize the system power consumption. ORBIT also provides a data processing library that includes two mechanisms namely adaptive delay-quality trade-off and data partitioning via multi-threading to optimize resource usage. Moreover, ORBIT supplies an annotation-based programming API for developers that significantly simplifies the application development and provides programming flexibility. Extensive microbenchmark evaluation and three case studies including seismic sensing, visual tracking using an ORBIT robot, and multi-camera 3D reconstruction, validate the generic design of ORBIT.

Fine-grained real-time metering is a fundamental service of wireless energy auditing networks, where metering data is transmitted from embedded wireless power meters to gateways for centralized processing, storage, and forwarding. Due to limited meter capability and wireless bandwidth, the increasing sampling rates and network scales needed to support new energy auditing applications pose significant challenges to metering data fidelity and secrecy . This article exploits the compression and encryption properties of compressive sensing (CS) to design a joint data compression and encryption (JICE) approach that addresses these two challenges simultaneously. Compared with a conventional signal processing pipeline that compresses and encrypts data sequentially, JICE reduces computation and space complexities due to its simple design. It thus leaves more processor time and available buffer space for handling lossy wireless transmissions. Moreover, JICE features an adaptive reconfiguration mechanism that selects the signal representation basis of CS at runtime among several candidate bases to achieve the best fidelity of the recovered data at the gateways. This mechanism enables JICE to adapt to changing power consumption patterns. On a smart plug platform, we implemented JICE and several baseline approaches including downsampling, lossless compression, and the pipeline approach. Extensive testbed experiments show that JICE achieves higher data delivery ratios and lower recovery distortions under a range of realistic settings. In particular, at a meter sampling rate of 8 Hz, JICE increases the number of meters supported by a gateway by 50%, compared with the commonly used pipeline approach, while keeping a signal distortion rate lower than 5%.

Appliance-level power usage monitoring may help conserve electricity in homes. Several existing systems achieve this goal by exploiting appliances’ power usage signatures identified in labor-intensive in situ training processes. Recent work shows that autonomous power usage monitoring can be achieved by supplementing a smart meter with distributed sensors that detect the working states of appliances. However, sensors must be carefully installed for each appliance, resulting in a high installation cost. This article presents Supero —the first ad hoc sensor system that can monitor appliance power usage without supervised training. By exploiting multisensor fusion and unsupervised machine learning algorithms, Supero can classify the appliance events of interest and autonomously associate measured power usage with the respective appliances. Our extensive evaluation in five real homes shows that Supero can estimate the energy consumption with errors less than 7.5%. Moreover, nonprofessional users can quickly deploy Supero with considerable flexibility.

Modern urban railways extensively use computerized-sensing and control technologies to achieve safe, reliable, and well-timed operations. However, the use of these technologies may provide a convenient leverage to cyber-attackers who have bypassed the air gaps and aim at causing safety incidents and service disruptions. In this paper, we study false data injection (FDI) attacks against railways' traction power systems (TPSes). Specifically, we analyze two types of FDI attacks on the train-borne voltage, current, and position sensor measurements -- which we call efficiency attack and safety attack -- that (i) maximize the system's total power consumption and (ii) mislead trains' local voltages to exceed given safety-critical thresholds, respectively. To counteract, we develop a global attack detection system that serializes a bad data detector anda novel secondary attack detector designed based on unique TPS characteristics. With intact position data of trains, our detection system can effectively detect the FDI attacks ontrains' voltage and current measurements even if the attacker has full and accurate knowledge of the TPS, attack detection, and real-time system state. Extensive simulations driven by realistic running profiles of trains verify that a TPS setup isvulnerable to the FDI attacks, but these attacks can be detected effectively by the proposed global monitoring.

Much recent work has applied existing fault detectors against attacks in cyber-physical control systems. The results demonstrate effectiveness in detecting simplistic attacks that cause fault-like disruptions. However, they do not address motivated and knowledgeable attackers who craft attacks using knowledge of the system including its method of detecting attacks. In this paper, we analyze the conditions for an attacker to bypass a dissipativity-theoretic fault detector adopted in the prior work. We show that the attacker can use a quadratic programming solver to efficiently compute false data injection attacks to bypass the detector. We show further that, by applying an OR gate to fuse binary detection results from a number of the detectors, with carefully chosen parameters, we can achieve an integrated detector bank that cannot be bypassed by an attacker, if the attacker can tamper with either the sensor or control data of the system. For an n-dimensional linear time-invariant system, the number of needed fault detectors is O(n!). This number can be dramatically reduced to O(n) under a realistic assumption that the system has converged before the attack starts. Simulations for voltage control based on an IEEE 39-bus power system model validate our analysis.

Desynchronized clocks among nodes in industrial Internet of Things (IoT) can degrade system performance and even lead to safety incidents. Clock synchronization protocols based on network message exchanges, though widely used in current industrial systems, are susceptible to delay attacks against the packet transmission. This vulnerability cannot be solved by conventional security measures such as encryption, and remains an open problem. This paper proposes to use the sine voltage waveform of a utility power grid to synchronize "things" connected to the same grid. Our experiments demonstrate that minute fluctuations of the voltage's cycle length encode fine-grained global time information in a city-scale utility grid. Based on this key result, we develop a clock synchronization approach that achieves sub-ms accuracy and is provably secure against packet delay attacks. Implementation results show that our approach achieves an average synchronization error of 0.1 ms between two IoT nodes that are 10 km apart. When the proposed system is deployed within the same floor of a building, the error reduces to 10 us.

Monitoring aquatic debris is of great interest to the ecosystems, marine life, human health, and water transport. This paper presents the design and implementation of SOAR-a vision-based surveillance robot system that integrates an off-the-shelf Android smartphone and a gliding robotic fish for debris monitoring in relatively calm waters. SOAR features real-time debris detection and coverage-based rotation scheduling algorithms. The image processing algorithms for debris detection are specifically designed to address the unique challenges in aquatic environments. The rotation scheduling algorithm provides effective coverage for sporadic debris arrivals despite camera's limited angular view. Moreover, SOAR is able to dynamically offload compute-intensive processing tasks to the cloud for battery power conservation. We have implemented a SOAR prototype and conducted extensive experimental evaluation. The results show that SOAR can accurately detect debris in the presence of various environment and system dynamics, and the rotation scheduling algorithm enables SOAR to capture debris arrivals with reduced energy consumption.

Monitoring aquatic environment is of great interest to the ecosystem, marine life, and human health. This article presents the design and implementation of Samba—an aquatic surveillance robot that integrates an off-the-shelf Android smartphone and a robotic fish to monitor harmful aquatic processes such as oil spills and harmful algal blooms. Using the built-in camera of the smartphone, Samba can detect spatially dispersed aquatic processes in dynamic and complex environment. To reduce the excessive false alarms caused by the nonwater area (e.g., trees on the shore), Samba segments the captured images and performs target detection in the identified water area only. However, a major challenge in the design of Samba is the high energy consumption resulted from continuous image segmentation. We propose a novel approach that leverages the power-efficient inertial sensors on smartphones to assist image processing. In particular, based on the learned mapping models between inertial and visual features, Samba uses real-time inertial sensor readings to estimate the visual features that guide image segmentation, significantly reducing the energy consumption and computation overhead. Samba also features a set of lightweight and robust computer vision algorithms, which detect harmful aquatic processes based on their distinctive color features. Last, Samba employs a feedback-based rotation control algorithm to adapt to spatiotemporal development of the target aquatic process. We have implemented a Samba prototype and evaluated it through extensive field experiments, lab experiments, and trace-driven simulations. The results show that Samba can achieve a 94% detection rate, a 5% false alarm rate, and a lifetime up to nearly 2 months.

Owing to the rich processing, multi-modal sensing, and versatile networking capabilities, smartphones are increasingly used to build data-intensive embedded sensing applications. However, various challenges must be systematically addressed before smartphones can be used as a generic embedded sensing platform, including high power consumption, lack of real-time functionality and user-friendly embedded programming support. This paper presents ORBIT, a smartphone-based platform for data-intensive embedded sensing applications. ORBIT features a tiered architecture, in which a smartphone can interface to an energy-efficient peripheral board and/or a cloud service. ORBIT as a platform addresses the shortcomings of current smartphones while utilizing their strengths. ORBIT provides a profile-based task partitioning allowing it to intelligently dispatch the processing tasks among the tiers to minimize the system power consumption. ORBIT also provides a data processing library that includes two mechanisms namely adaptive delay/quality trade-off and data partitioning via multi-threading to optimize resource usage. Moreover, ORBIT supplies an annotation based programming API for developers that significantly simplifies the application development and provides programming flexibility. Extensive microbenchmark evaluation and two case studies including seismic sensing and multi-camera 3D reconstruction, validate the generic design of ORBIT.

Monitoring aquatic environment is of great interest to the ecosystem, marine life, and human health. This paper presents the design and implementation of Samba -- an aquatic surveillance robot that integrates an off-the-shelf Android smartphone and a robotic fish to monitor harmful aquatic processes such as oil spill and harmful algal blooms. Using the built-in camera of on-board smartphone, Samba can detect spatially dispersed aquatic processes in dynamic and complex environments. To reduce the excessive false alarms caused by the non-water area (e.g., trees on the shore), Samba segments the captured images and performs target detection in the identified water area only. However, a major challenge in the design of Samba is the high energy consumption resulted from the continuous image segmentation. We propose a novel approach that leverages the power-efficient inertial sensors on smartphone to assist the image processing. In particular, based on the learned mapping models between inertial and visual features, Samba uses real-time inertial sensor readings to estimate the visual features that guide the image segmentation, significantly reducing energy consumption and computation overhead. Samba also features a set of lightweight and robust computer vision algorithms, which detect harmful aquatic processes based on their distinctive color features. Lastly, Samba employs a feedback-based rotation control algorithm to adapt to spatiotemporal evolution of the target aquatic process. We have implemented a Samba prototype and evaluated it through extensive field experiments, lab experiments, and trace-driven simulations. The results show that Samba can achieve 94% detection rate, 5% false alarm rate, and a lifetime up to nearly two months.

Fine-grained real-time metering is a fundamental service of wireless energy auditing networks, where metering data is transmitted from embedded power meters to gateways for centralized processing, storage, and forwarding. Due to limited meter capability and wireless bandwidth, the increasing sampling rates and network scales needed to support new energy auditing applications pose significant challenges to metering data fidelity and secrecy. This paper exploits the compression and encryption properties of compressive sensing (CS) to design a joint data compression and encryption (JICE) approach that addresses these two challenges simultaneously. Compared with a conventional signal processing pipeline that compresses and encrypts data sequentially, JICE reduces computation and storage complexities due to its simple design. It thus leaves more processor time and available buffer space for handling lossy wireless transmissions. Moreover, JICE features a machine-learning-based reconfiguration mechanism that adapts its signal representation basis to changing power patterns autonomously. On a smart plug platform, we implemented JICE and several baseline approaches including downsampling, lossless compression, and the pipeline approach. Extensive testbed experiments show that JICE achieves higher data delivery ratios and lower recovery distortions under a range of realistic settings. In particular, JICE increases the number of meters supported by a gateway by 50%, compared with the pipeline approach, while keeping a distortion rate lower than 5%.

Modern information and communication technologies used by electric power grids are subject to cyber-security threats. This article studies the impact of integrity attacks on real-time pricing (RTP), an emerging feature of advanced power grids that can improve system efficiency. Recent studies have shown that RTP creates a closed loop formed by the mutually dependent real-time price signals and price-taking demand. Such a closed loop can be exploited by an adversary whose objective is to destabilize the pricing system. Specifically, small malicious modifications to the price signals can be iteratively amplified by the closed loop, causing highly volatile prices, fluctuating power demand, and increased system operating cost. This article adopts a control-theoretic approach to deriving the fundamental conditions of RTP stability under basic demand, supply, and RTP models that characterize the essential behaviors of consumers, suppliers, and system operators, as well as two broad classes of integrity attacks, namely, the scaling and delay attacks. We show that, under an approximated linear time-invariant formulation, the RTP system is at risk of being destabilized only if the adversary can compromise the price signals advertised to consumers, by either reducing their values in the scaling attack or providing old prices to over half of all consumers in the delay attack. The results provide useful guidelines for system operators to analyze the impact of various attack parameters on system stability so that they may take adequate measures to secure RTP systems.

Data centers have become a critical computing infrastructure in the era of cloud computing. Temperature monitoring and forecasting are essential for preventing server shutdowns because of overheating and improving a data center’s energy efficiency. This article presents a novel cyber-physical approach for temperature forecasting in data centers, one that integrates Computational Fluid Dynamics (CFD) modeling, in situ wireless sensing, and real-time data-driven prediction. To ensure forecasting fidelity, we leverage the realistic physical thermodynamic models of CFD to generate transient temperature distribution and calibrate it using sensor feedback. Both simulated temperature distribution and sensor measurements are then used to train a real-time prediction algorithm. As a result, our approach reduces not only the computational complexity of online temperature modeling and prediction, but also the number of deployed sensors, which enables a portable, noninvasive thermal monitoring solution that does not rely on the infrastructure of a monitored data center. We extensively evaluated the proposed system on a rack of 15 servers and a testbed of five racks and 229 servers in a small-scale production data center. Our results show that our system can predict the temperature evolution of servers with highly dynamic workloads at an average error of 0.52○C, within a duration up to 10 minutes. Moreover, our approach can reduce the required number of sensors by 67% while maintaining desirable prediction fidelity.

Current data centers often adopt conservative and static settings for cooling and air circulation systems, leading to excessive energy consumption. This paper presents the design and evaluation of PTEC -- a system for predictive thermal and energy control in data centers. PTEC leverages the server built-in sensors and monitoring utilities, as well as a wireless sensor network, to monitor both the cyber and physical status of a data center. By predicting the temperature evolution of a data center in real time, PTEC finds the temperature set points, the cold air supply rates, and the speeds of server internal fans to minimize the expected total energy consumption of cooling and circulation systems. Moreover, PTEC enforces the upper bounds on server inlet temperatures and their temporal variations to prevent server overheating and reduce server hardware failure rate. We evaluated PTEC on a hardware test bed consisting of 15 servers and a total of 23 temperature and power sensors, as well as through Computational Fluid Dynamics (CFD) simulations based on real data traces collected from a data center with 229 servers. The experimental results show that PTEC can reduce the cooling and circulation energy consumption by more than 30%, compared with baseline thermal control strategies.

Water resources and aquatic ecosystems are facing increasing threats from climate change, improper waste disposal, and oil spill incidents. It is of great interest to deploy mobile sensors to detect and monitor certain diffusion processes (e.g., chemical pollutants) that are harmful to aquatic environments. In this paper, we propose an accuracy-aware diffusion process profiling approach using smart aquatic mobile sensors such as robotic fish. In our approach, the robotic sensors collaboratively profile the characteristics of a diffusion process including source location, discharged substance amount, and its evolution over time. In particular, the robotic sensors reposition themselves to progressively improve the profiling accuracy. We formulate a novel movement scheduling problem that aims to maximize the profiling accuracy subject to the limited sensor mobility and energy budget. We develop an efficient greedy algorithm and a more complex near-optimal radial algorithm to solve the problem. We conduct extensive simulations based on real data traces of GPS localization errors, robotic fish movement, and wireless communication. The results show that our approach can accurately profile dynamic diffusion processes under tight energy budgets. Moreover, a preliminary evaluation based on the implementation on TelosB motes validates the feasibility of deploying our profiling algorithms on mote-class robotic sensor platforms.

Monitoring important aquatic processes like harmful algal blooms is of increasing interest to public health, ecosystem sustainability, marine biology, and aquaculture industry. This article presents a novel approach to spatiotemporal aquatic field reconstruction using inexpensive, low-power mobile sensing platforms called robotic fish . Robotic fish networks are a typical example of cyber-physical systems where the design of cyber components (sensing, communication, and information processing) must account for inherent physical dynamics of the robots and the aquatic environment. Our approach features a rendezvous-based mobility control scheme where robotic fish collaborate in the form of a swarm to sense the aquatic environment in a series of carefully chosen rendezvous regions. We design a novel feedback control algorithm that maintains the desirable level of wireless connectivity for a sensor swarm in the presence of significant environment and system dynamics. Information-theoretic analysis is used to guide the selection of rendezvous regions so that the spatiotemporal field reconstruction accuracy is maximized subject to the limited sensor mobility. The effectiveness of our approach is validated via implementation on sensor hardware and extensive simulations based on real data traces of water surface temperature field and on-water ZigBee wireless communication.

Modern information and communication technologies used by smart grids are subject to cybersecurity threats. This paper studies the impact of integrity attacks on real-time pricing (RTP), a key feature of smart grids that uses such technologies to improve system efficiency. Recent studies have shown that RTP creates a closed loop formed by the mutually dependent real-time price signals and price-taking demand. Such a closed loop can be exploited by an adversary whose objective is to destabilize the pricing system. Specifically, small malicious modifications to the price signals can be iteratively amplified by the closed loop, causing inefficiency and even severe failures such as blackouts. This paper adopts a control-theoretic approach to deriving the fundamental conditions of RTP stability under two broad classes of integrity attacks, namely, the scaling and delay attacks. We show that the RTP system is at risk of being destabilized only if the adversary can compromise the price signals advertised to smart meters by reducing their values in the scaling attack, or by providing old prices to over half of all consumers in the delay attack. The results provide useful guidelines for system operators to analyze the impact of various attack parameters on system stability, so that they may take adequate measures to secure RTP systems.

Recent years have witnessed pilot deployments of inexpensive wireless sensor networks (WSNs) for active volcano monitoring. This paper studies the problem of picking arrival times of primary waves (i.e., P-phases) received by seismic sensors, one of the most critical tasks in volcano monitoring. Two fundamental challenges must be addressed. First, it is virtually impossible to download the real-time high-frequency seismic data to a central station for P-phase picking due to limited wireless network bandwidth. Second, accurate P-phase picking is inherently computation-intensive, and is thus prohibitive for many low-power sensor platforms. To address these challenges, we propose a new P-phase picking approach for hierarchical volcano monitoring WSNs where a large number of inexpensive sensors are used to collect fine-grained, real-time seismic signals while a small number of powerful coordinator nodes process collected data and pick accurate P-phases. We develop a suite of new in-network signal processing algorithms for accurate P-phase picking, including lightweight signal pre-processing at sensors, sensor selection at coordinators as well as signal compression and reconstruction algorithms. Testbed experiments and extensive simulations based on real data collected from a volcano show that our approach achieves accurate P-phase picking while only 16% of the sensor data are transmitted.

In this paper we advocate the use of workflow---describing how a system provides its intended functionality---as a pillar of cybersecurity analysis and propose a holistic workflow-oriented assessment framework. While workflow models are currently used in the area of performance and reliability assessment, these approaches are designed neither to assess a system in the presence of an active attacker, nor to assess security aspects such as confidentiality. On the other hand, existing security assessment methods typically focus on modeling the active attacker (e.g., attack graphs), but many rely on restrictive models that are not readily applicable to complex (e.g., cyber-physical or cyber-human) systems.

As a key technology of home area networks in smart grids, fine-grained power usage monitoring may help conserve electricity. Several existing systems achieve this goal by exploiting appliances' power usage signatures identified in labor-intensive in situ training processes. Recent work shows that autonomous power usage monitoring can be achieved by supplementing a smart meter with distributed sensors that detect the working states of appliances. However, sensors must be carefully installed for each appliance, resulting in high installation cost. This paper presents Supero - the first ad hoc sensor system that can monitor appliance power usage without supervised training. By exploiting multisensor fusion and unsupervised machine learning algorithms, Supero can classify the appliance events of interest and autonomously associate measured power usage with the respective appliances. Our extensive evaluation in five real homes shows that Supero can estimate the energy consumption with errors less than 7.5%. Moreover, non-professional users can quickly deploy Supero with considerable flexibility.

Volcano monitoring is of great interest to public safety and scientific explorations. However, traditional volcanic instrumentation such as broadband seismometers are expensive, power hungry, bulky, and difficult to install. Wireless sensor networks (WSNs) offer the potential to monitor volcanoes on unprecedented spatial and temporal scales. However, current volcanic WSN systems often yield poor monitoring quality due to the limited sensing capability of low-cost sensors and unpredictable dynamics of volcanic activities. In this article, we propose a novel quality-driven approach to achieving real-time, distributed, and long-lived volcanic earthquake detection and timing. By employing novel in-network collaborative signal processing algorithms, our approach can meet stringent requirements on sensing quality (i.e., low false alarm/missing rate, short detection delay, and precise earthquake onset time) at low power consumption. We have implemented our algorithms in TinyOS and conducted extensive evaluation on a testbed of 24 TelosB motes as well as simulations based on real data traces collected during 5.5 months on an active volcano. We show that our approach yields near-zero false alarm/missing rate, less than one second of detection delay, and millisecond precision earthquake onset time while achieving up to six-fold energy reduction over the current data collection approach.

Wireless sensor networks are typically composed of low-cost sensors that are deeply integrated in physical environments. As a result, the sensing performance of a wireless sensor network is inevitably undermined by biases in imperfect sensor hardware and the noises in data measurements. Although a variety of calibration methods have been proposed to address these issues, they often adopt the device-level approach that becomes intractable for moderate-to large-scale networks. In this article, we propose a two-tier system-level calibration approach for a class of sensor networks that employ data fusion to improve the sensing performance. In the first tier of our calibration approach, each sensor learns its local sensing model from noisy measurements using an online algorithm and only transmits a few model parameters. In the second tier, sensors' local sensing models are then calibrated to a common system sensing model. Our approach fairly distributes computation overhead among sensors and significantly reduces the communication overhead of calibration compared with the device-level approach. Based on this approach, we develop an optimal model calibration scheme that maximizes the target detection probability of a sensor network under bounded false alarm rate. Our approach is evaluated by both experiments on a testbed of TelosB motes and extensive simulations based on synthetic datasets as well as data traces collected in a real vehicle detection experiment. The results demonstrate that our system-level calibration approach can significantly boost the detection performance of sensor networks in scenarios with low signal-to-noise ratios.

Water resources and aquatic ecosystems are facing increasing threats from climate change, improper waste disposal, and oil spill incidents. It is of great interest to deploy mobile sensors to detect and monitor certain diffusion processes (e.g., chemical pollutants) that are harmful to aquatic environments. In this paper, we propose an accuracy-aware diffusion process profiling approach using smart aquatic mobile sensors such as robotic fish. In our approach, the robotic sensors collaboratively profile the characteristics of a diffusion process including source location, discharged substance amount, and its evolution over time. In particular, the robotic sensors reposition themselves to progressively improve the profiling accuracy. We formulate a novel movement scheduling problem that aims to maximize the profiling accuracy subject to limited sensor mobility and energy budget. We develop an efficient greedy algorithm and a more complex near-optimal radial algorithm to solve the problem. We conduct extensive simulations based on real data traces of robotic fish movement and wireless communication. The results show that our approach can accurately profile dynamic diffusion processes under tight energy budgets. Moreover, a preliminary evaluation based on the implementation on TelosB motes validates the feasibility of deploying our movement scheduling algorithms on mote-class robotic sensor platforms.

Data centers have become a critical computing infrastructure in the era of cloud computing. Temperature monitoring and forecasting are essential for preventing overheating-induced server shutdowns and improving a data center's energy efficiency. This paper presents a novel cyber-physical approach for temperature forecasting in data centers, which integrates Computational Fluid Dynamics (CFD) modeling, in situ wireless sensing, and real-time data-driven prediction. To ensure the forecasting fidelity, we leverage the realistic physical thermodynamic models of CFD to generate transient temperature distribution and calibrate it using sensor feedback. Both simulated temperature distribution and sensor measurements are then used to train a real-time prediction algorithm. As a result, our approach significantly reduces the computational complexity of online temperature modeling and prediction, which enables a portable, noninvasive thermal monitoring solution that does not rely on the infrastructure of monitored data center. We extensively evaluated our system on a rack of 15 servers and a test bed of five racks and 229 servers in a production data center. Our results show that our system can predict the temperature evolution of servers with highly dynamic workloads at an average error of 0.52C, within a duration up to 10 minutes.

Monitoring important aquatic processes like harmful algal blooms is of increasing interest to public health, ecosystem sustainability, marine biology, and aquaculture industry. This paper presents a novel approach to spatiotemporal aquatic field reconstruction using inexpensive, low-power, mobile sensing platforms called robotic fish. Robotic fish networks are a typical example of Cyber-Physical Systems where the design of cyber components (sensing, communication, and information processing) must account for inherent physical dynamics of the robots and the aquatic environment. Our approach features a rendezvous-based mobility control scheme where robotic fish collaborate in the form of a swarm to sense the aquatic environment in a series of carefully chosen rendezvous regions. We design a novel feedback control algorithm that maintains the desirable level of wireless connectivity for a sensor swarm in the presence of significant environment and system dynamics. Information-theoretic analysis is used to guide the selection of rendezvous regions so that the spatiotemporal field reconstruction accuracy is maximized subject to the limited sensor mobility. The effectiveness of our approach is validated via implementation on sensor hardware and extensive simulations based on real data traces of water surface temperature field and on-water ZigBee wireless communication.

Many Cyber-Physical Systems (CPS) are composed of low-cost devices that are deeply integrated with physical environments. As a result, the performance of a CPS system is inevitably undermined by various physical uncertainties , which include stochastic noises, hardware biases, unpredictable environment changes, and dynamics of the physical process of interest. Traditional solutions to these issues (e.g., device calibration and collaborative signal processing) work in an open-loop fashion and hence often fail to adapt to the uncertainties after system deployment. In this article, we propose an adaptive system-level calibration approach for a class of CPS systems whose primary objective is to detect events or targets of interest. Through collaborative data fusion, our calibration approach features a feedback control loop that exploits system heterogeneity to mitigate the impact of aforementioned uncertainties on the system performance. In contrast to existing heuristic-based solutions, our control-theoretical calibration algorithm can ensure provable system stability and convergence. We also develop a routing algorithm for fusion-based multihop CPS systems that is robust to communication unreliability and delay. Our approach is evaluated by both experiments on a testbed of Tmotes as well as extensive simulations based on data traces gathered from a real vehicle detection experiment. The results demonstrate that our calibration algorithm enables a CPS system to maintain the optimal sensing performance in the presence of various system and environmental dynamics.

Wireless sensor networks (WSNs) have been increasingly available for critical applications such as security surveillance and environmental monitoring. An important performance measure of such applications is sensing coverage that characterizes how well a sensing field is monitored by a network. Although advanced collaborative signal processing algorithms have been adopted by many existing WSNs, most previous analytical studies on sensing coverage are conducted based on overly simplistic sensing models (e.g., the disc model) that do not capture the stochastic nature of sensing. In this paper, we attempt to bridge this gap by exploring the fundamental limits of coverage based on stochastic data fusion models that fuse noisy measurements of multiple sensors. We derive the scaling laws between coverage, network density, and signal-to-noise ratio (SNR). We show that data fusion can significantly improve sensing coverage by exploiting the collaboration among sensors when several physical properties of the target signal are known. In particular, for signal path loss exponent of (typically between 2.0 and 5.0), ρf= O(ρd1-1/k, where ρfand ρdare the densities of uniformly deployed sensors that achieve full coverage under the fusion and disc models, respectively. Moreover, data fusion can also reduce network density for regularly deployed networks and mobile networks where mobile sensors can relocate to fill coverage holes. Our results help understand the limitations of the previous analytical results based on the disc model and provide key insights into the design of WSNs that adopt data fusion algorithms. Our analyses are verified through extensive simulations based on both synthetic data sets and data traces collected in a real deployment for vehicle detection.

Recent years have seen the growing deployments of Cyber-Physical Systems (CPSs) in many mission-critical applications such as security, civil infrastructure, and transportation. These applications often impose stringent requirements on system sensing fidelity and timeliness. However, existing approaches treat these two concerns in isolation and hence are not suitable for CPSs where system fidelity and timeliness are dependent on each other because of the tight integration of computational and physical resources. In this paper, we propose a holistic approach called Fidelity-Aware Utilization Controller (FAUC) for Wireless Cyber-physical Surveillance (WCS) systems that combine low-end sensors with cameras for large-scale ad hoc surveillance in unplanned environments. By integrating data fusion with feedback control, FAUC can enforce a CPU utilization upper bound to ensure the system's real-time schedulability although CPU workloads vary significantly at runtime because of stochastic detection results. At the same time, FAUC optimizes system fidelity and adjusts the control objective of CPU utilization adaptively in the presence of variations of target/noise characteristics. We have implemented FAUC on a small-scale WCS testbed consisting of TelosB/Iris motes and cameras. Moreover, we conduct extensive simulations based on real acoustic data traces collected in a vehicle surveillance experiment. The testbed experiments and the trace-driven simulations show that FAUC can achieve robust fidelity and real-time guarantees in dynamic environments.

Mission-critical target detection imposes stringent performance requirements for wireless sensor networks, such as high detection probabilities and low false alarm rates. Data fusion has been shown as an effective technique for improving system detection performance by enabling efficient collaboration among sensors with limited sensing capability. Due to the high cost of network deployment, it is desirable to place sensors at optimal locations to achieve maximum detection performance. However, for sensor networks employing data fusion, optimal sensor placement is a nonlinear and nonconvex optimization problem with prohibitively high computational complexity. In this paper, we present fast sensor placement algorithms based on a probabilistic data fusion model. Simulation results show that our algorithms can meet the desired detection performance with a small number of sensors while achieving up to seven-fold speedup over the optimal algorithm.

Real-time detection is an important requirement of many mission-critical wireless sensor network applications such as battlefield monitoring and security surveillance. Due to the high network deployment cost, it is crucial to understand and predict the real-time detection capability of a sensor network. However, most existing real-time analyses are based on overly simplistic sensing models (e.g., the disc model) that do not capture the stochastic nature of detection. In practice, data fusion has been adopted in a number of sensor systems to deal with sensing uncertainty and enable efficient collaboration among resource-limited sensors. However, real-time performance analysis of sensor networks designed based on data fusion has received little attention. In this paper, we bridge this gap by investigating the fundamental real-time detection performance of large-scale sensor networks under stochastic sensing models. In particular, we consider two basic data fusion schemes, i.e., value fusion and decision fusion. Our results show that data fusion is effective in achieving stringent performance requirements such as short detection delay and low false alarm rates. Moreover, value fusion and decision fusion are suitable for low and high signal-to-noise ratio scenarios, respectively. Our results help understand the impact of data fusion and provide important guidelines for the design of real-time wireless sensor networks for intrusion detection. Our analyses are verified through extensive simulations based on both synthetic data sets and data traces collected in a real deployment for vehicle detection. The results show that data fusion can reduce the network density by about 60 percent compared with the disc model while detecting any intruder within one detection period at a false alarm rate lower than five percent.

Wireless sensor networks (WSNs) are typically composed of low-cost sensors that are deeply integrated with physical environments. As a result, the sensing performance of a WSN is inevitably undermined by various physical uncertainties, which include stochastic sensor noises, unpredictable environment changes and dynamics of the monitored phenomenon. Traditional solutions (e.g., sensor calibration and collaborative signal processing) work in an open-loop fashion and hence fail to adapt to these uncertainties after system deployment. In this paper, we propose an adaptive system-level calibration approach for a class of sensor networks that employ data fusion to improve system sensing performance. Our approach features a feedback control loop that exploits sensor heterogeneity to deal with the aforementioned uncertainties in calibrating system performance. In contrast to existing heuristic based solutions, our control-theoretical calibration algorithm can ensure provable system stability and convergence. We also systematically analyze the impacts of communication reliability and delay, and propose an optimal routing algorithm that minimizes the impact of packet loss on system stability. Our approach is evaluated by both experiments on a testbed of Tmotes as well as extensive simulations based on data traces gathered from a real vehicle detection experiment. The results demonstrate that our calibration algorithm enables a network to maintain the optimal detection performance in the presence of various system and environmental dynamics.

Recent years have seen the growing deployments of Cyber-Physical Systems (CPSs) in many mission-critical applications such as security, civil infrastructure, and transportation. These applications often impose stringent requirements on system sensing fidelity and timeliness. However, existing approaches treat these two concerns in isolation and hence are not suitable for CPSs where system fidelity and timeliness are dependent of each other because of the tight integration of computational and physical resources. In this paper, we propose a holistic approach called Fidelity-Aware Utilization Controller (FAUC) for Wireless Cyber-physical Surveillance (WCS) systems that combine low-end sensors with cameras for large-scale ad hoc surveillance in unplanned environments. By integrating data fusion with feedback control, FAUC can enforce a CPU utilization upper bound to ensure the system's real-time schedulability although CPU workloads vary significantly at runtime because of stochastic detection results. At the same time, FAUC optimizes system fidelity and adjusts the control objective of CPU utilization adaptively in the presence of variations of target/noise characteristics. We have implemented FAUC on a small-scale WCS testbed consisting of TelosB/Iris motes and cameras. Our extensive experiments on light and acoustic target detection show that FAUC can achieve robust fidelity and real-time guarantees in dynamic environments.

Volcano monitoring is of great interest to public safety and scientific explorations. However, traditional volcanic instrumentation such as broadband seismometers are expensive, power-hungry, bulky, and difficult to install. Wireless sensor networks (WSNs) offer the potential to monitor volcanoes at unprecedented spatial and temporal scales. However, current volcanic WSN systems often yield poor monitoring quality due to the limited sensing capability of low-cost sensors and unpredictable dynamics of volcanic activities. Moreover, they are designed only for short-term monitoring due to the high energy consumption of centralized data collection. In this paper, we propose a novel quality-driven approach to achieving real-time, in-situ, and long-lived volcanic earthquake detection. By employing novel in-network collaborative signal processing algorithms, our approach can meet stringent requirements on sensing quality (low false alarm/missing rate and precise earthquake onset time) at low power consumption. We have implemented our algorithms in TinyOS and conducted extensive evaluation on a testbed of 24 TelosB motes as well as simulations based on real data traces collected during 5.5 months on an active volcano. We show that our approach yields near-zero false alarm/missing rate and less than one second of detection delay while achieving up to 6-fold energy reduction over the current data collection approach.

Wireless sensor networks are typically composed of low-cost sensors that are deeply integrated in physical environments. As a result, the sensing performance of a wireless sensor network is inevitably undermined by biases in imperfect sensor hardware and the noises in data measurements. Although a variety of calibration methods have been proposed to address these issues, they often adopt the device-level approach that becomes intractable for moderate- to large-scale networks. In this paper, we propose a two-tier system-level calibration approach for a class of sensor networks that employ data fusion to improve the sensing performance. In the first tier of our calibration approach, each sensor learns its local sensing model from noisy measurements using an online algorithm and only transmits a few model parameters. In the second tier, sensors' local sensing models are then calibrated to a common system sensing model. Our approach fairly distributes computation overhead among sensors and significantly reduces the communication overhead of calibration. Based on this approach, we develop an optimal model calibration scheme that maximizes the target detection probability of a sensor network under bounded false alarm rate. Our approach is evaluated by both experiments on a testbed of TelosB motes and extensive simulations based on data traces collected in a real vehicle detection experiment. The results demonstrate that our system-level calibration approach can significantly boost the detection performance of sensor networks in the scenarios with low signal-to-noise ratios.

Recent years have witnessed the deployments of wireless sensor networks in a class of mission-critical applications such as object detection and tracking. These applications often impose stringent Quality-of-Service requirements including high detection probability, low false alarm rate, and bounded detection delay. Although a dense all-static network may initially meet these Quality-of-Service requirements, it does not adapt to unpredictable dynamics in network conditions (e.g., coverage holes caused by death of nodes) or physical environments (e.g., changed spatial distribution of events). This paper exploits reactive mobility to improve the target detection performance of wireless sensor networks. In our approach, mobile sensors collaborate with static sensors and move reactively to achieve the required detection performance. Specifically, mobile sensors initially remain stationary and are directed to move toward a possible target only when a detection consensus is reached by a group of sensors. The accuracy of final detection result is then improved as the measurements of mobile sensors have higher Signal-to-Noise Ratios after the movement. We develop a sensor movement scheduling algorithm that achieves near-optimal system detection performance under a given detection delay bound. The effectiveness of our approach is validated by extensive simulations using the real data traces collected by 23 sensor nodes.

Wireless sensor networks (WSNs) deployed for mission-critical applications face the fundamental challenge of meeting stringent spatiotemporal performance requirements using nodes with limited sensing capacity. Although advance network planning and dense node deployment may initially achieve the required performance, they often fail to adapt to the unpredictability and variability of physical reality. This paper explores efficient use of mobile sensors to address limitations of static WSNs for target detection. We propose a data-fusion-based detection model that enables static and mobile sensors to effectively collaborate in target detection. An optimal sensor movement scheduling algorithm is developed to minimize the total moving distance of sensors while achieving a set of spatiotemporal performance requirements including high detection probability, low system false alarm rate, and bounded detection delay. The effectiveness of our approach is validated by extensive simulations based on real data traces collected by 23 sensor nodes.

Wireless sensor networks (WSNs) have been increasingly available for critical applications such as security surveil-lance and environmental monitoring. An important per-formance measure of such applications is sensing coverage that characterizes how well a sensing field is monitored by a network. Although advanced collaborative signal process-ing algorithms have been adopted by many existing WSNs, most previous analytical studies on sensing coverage are con-ducted based on overly simplistic sensing models (e.g., the disc model) that do not capture the stochastic nature of sens-ing. In this paper, we attempt to bridge this gap by explor-ing the fundamental limits of coverage based on stochastic data fusion models that fuse noisy measurements of multi-ple sensors. We derive the scaling laws between coverage, network density, and signal-to-noise ratio (SNR). We show that data fusion can significantly improve sensing coverage by exploiting the collaboration among sensors. In particu-lar, for signal path loss exponent of k (typically between 2.0 and 5.0), ρf = O(ρ1−1/kd), where ρf and ρd are the densi-ties of uniformly deployed sensors that achieve full coverage under the fusion and disc models, respectively. Our results help understand the limitations of the previous analytical re-sults based on the disc model and provide key insights into the design of WSNs that adopt data fusion algorithms. Our analyses are verified through extensive simulations based on both synthetic data sets and data traces collected in a real deployment for vehicle detection.

Real-time detection is an important requirement of many mission-critical wireless sensor network applications such as battlefield monitoring and security surveillance. Due to the high network deployment cost, it is crucial to understand and predict the real-time detection capability of a sensor network. However, most existing real-time analyses are based on overly simplistic sensing models (e.g., the disc model) that do not capture the stochastic nature of detection. In practice, data fusion has been adopted in a number of sensor systems to deal with sensing uncertainty and enable the collaboration among sensors. However, real-time performance analysis of sensor networks designed based on data fusion has received little attention. In this paper, we bridge this gap by investigating the fundamental real-time detection performance of large-scale sensor networks under stochastic sensing models. Our results show that data fusion is effective in achieving stringent performance requirements such as short detection delay and low false alarm rates, especially in the scenarios with low signal-to-noise ratios (SNRs). Data fusion can reduce the network density by about 60% compared with the disc model while detecting any intruder within one detection period at a false alarm rate lower than 2%. In contrast, the disc model is only suitable when the SNR is sufficiently high. Our results help understand the impact of data fusion and provide important guidelines for the design of real-time wireless sensor networks for intrusion detection.

Recent years have witnessed the deployments of wireless sensor networks in a class of mission-critical applications such as object detection and tracking. These applications often impose stringent QoS requirements including high detection probability, low false alarm rate and bounded detection delay. Although a dense all-static network may initially meet these QoS requirements, it does not adapt to unpredictable dynamics in network conditions (e.g., coverage holes caused by death of nodes) or physical environments (e.g., changed spatial distribution of events). This paper exploits reactive mobility to improve the target detection performance of wireless sensor networks. In our approach, mobile sensors collaborate with static sensors and move reactively to achieve the required detection performance. Specifically, mobile sensors initially remain stationary and are directed to move toward a possible target only when a detection consensus is reached by a group of sensors. The accuracy of final detection result is then improved as the measurements of mobile sensors have higher signal-to-noise ratios after the movement. We develop a sensor movement scheduling algorithm that achieves near-optimal system detection performance within a given detection delay bound. The effectiveness of our approach is validated by extensive simulations using the real data traces collected by 23 sensor nodes.

Mission-critical target detection imposes stringent performance requirements for wireless sensor networks, such as high detection probabilities and low false alarm rates. Data fusion has been shown as an effective technique for improving system detection performance by enabling efficient collaboration among sensors with limited sensing capability. Due to the high cost of network deployment, it is desirable to place sensors at optimal locations to achieve maximum detection performance. However, for sensor networks employing data fusion, optimal sensor placement is a non-linear optimizationproblem with prohibitive computational complexity. In this paper, we present fast sensor placement algorithms based on a probabilistic data fusion model.Simulation results show that our algorithms can meet the desired detection performance with a small number of sensors while achieving up to 7-fold speedup over the optimal algorithm.