Multimodal human action recognition (HAR) utilizes complementary data for activity classification. Built on traditional HAR tasks, recent advances in Large Language Models (LLMs) enable detailed descriptions and causal reasoning of human actions, advancing new tasks of human action understanding (HAU) and human action reasoning (HARn). However, most LLMs, especially multimodal Large Vision-Language Models (LVLMs), struggle with modalities other than RGB images, like depth, IMU, ormmWave, due to a lack of large-scale datasets in these task domains. Existing HAR datasets provide only coarse-grained annotations, in-sufficient for depicting the detailed action dynamics required in HAU and HARn tasks. Simply combining annotations and generating captions with LLMs often lacks necessary logical and spatiotemporal consistency. In this paper, we introduce CUHK-X, a large-scale multi-modal dataset and benchmarks for HAR, HAU, and HARn. It includes 64,267 samples of 40 actions performed by 30 participants across two indoor environments, covering diverse daily scenarios. To address the challenge of spatiotemporal inconsistencies in captions, we propose a prompt-based scene creation method that leverages LLMs to generate logically connected activity sequences. CUHK-X also includes three benchmarks with six tasks to evaluate state-of-the-art models. Experimental results show average accuracies of 76.52% for HAR, 40.76% for HAU, and 70.25% for HARn. This large-scale multimodal dataset aims to empower the research community to apply, develop, and adapt data-intensive learning techniques for a wide range of human activity-related tasks.

The prevailing single-source paradigm in wireless sensing produces specialized models that are unscalable and generalize poorly to new tasks. Multi-source pre-training offers a path toward a generalist backbone but poses challenges including task heterogeneity, data redundancy, structural incompatibility, and the lack of a general-purpose pre-training objective. To address these issues, we propose WiSwiss, a comprehensive self-supervised multi-source pre-training framework that learns a general-purpose backbone for each modality. WiSwiss integrates semantic deduplication for dataset curation and a transformation-invariant pre-training objective. Experiments show that WiSwiss outperforms models trained from scratch, improving WiFi and mmWave performance by 4.5% and 10.3%, respectively, while reducing fine-tuning data requirements by 22.2% and 28.6%. We also present a qualitative study of scaling laws, showing that gains are task-dependent and that larger models require sufficiently large and diverse pre-training corpora to achieve substantial improvements.

Federated Learning (FL) can be affected by data and device heterogeneities, caused by clients' different local data distributions and latencies in uploading model updates (i.e., staleness). Traditional schemes consider these heterogeneities as two separate and independent aspects, but this assumption is unrealistic in practical FL scenarios where these heterogeneities are intertwined. In these cases, traditional FL schemes are ineffective, and a better approach is to convert a stale model update into a unstale one. In this paper, we present a new FL framework that ensures the accuracy and computational efficiency of this conversion, hence effectively tackling the intertwined heterogeneities that may cause unlimited staleness in model updates. Our basic idea is to estimate the distributions of clients' local training data from their uploaded stale model updates, and use these estimations to compute unstale client model updates. In this way, our approach does not require any auxiliary dataset nor the clients' local models to be fully trained, and does not incur any additional computation or communication overhead at client devices. We compared our approach with the existing FL strategies on mainstream datasets and models, and showed that our approach can improve the trained model accuracy by up to 25% and reduce the number of required training epochs by up to 35%. Source codes can be found at: https://github.com/pittisl/FL-with-intertwined-heterogeneity.

Text-to-video (T2V) generation has been recently enabled by transformer-based diffusion models, but current T2V models lack capabilities in adhering to the real-world common knowledge and physical rules, due to their limited understanding of physical realism and deficiency in temporal modeling. Existing solutions are either data-driven or require extra model inputs, but cannot be generalizable to out-of-distribution domains. In this paper, we present PhyT2V, a new data-independent T2V technique that expands the current T2V model’s capability of video generation to out-of-distribution domains, by enabling chain-of-thought and step-back reasoning in T2V prompting. Our experiments show that PhyT2V improves existing T2V models’ adherence to real-world physical rules by 2.3x, and achieves 35% improvement compared to T2V prompt enhancers.

Prosthetic legs play a pivotal role in clinical rehabilitation, allowing individuals with lower-limb amputations the ability to regain mobility and improve their quality of life. Gait analysis is fundamental for optimizing prosthesis design and alignment, directly impacting the mobility and life quality of individuals with lower-limb amputations. Vision-based machine learning (ML) methods offer a scalable and non-invasive solution to gait analysis, but face challenges in correctly detecting and analyzing prosthesis, due to their unique appearances and new movement patterns. In this paper, we aim to bridge this gap by introducing a multi-purpose dataset, namely ProGait, to support multiple vision tasks including Video Object Segmentation, 2D Human Pose Estimation, and Gait Analysis (GA). ProGait provides 412 video clips from four above-knee amputees when testing multiple newly-fitted prosthetic legs through walking trials, and depicts the presence, contours, poses, and gait patterns of human subjects with transfemoral prosthetic legs. Alongside the dataset itself, we also present benchmark tasks and fine-tuned baseline models to illustrate the practical application and performance of the ProGait dataset. We compared our baseline models against pre-trained vision models, demonstrating improved generalizability when applying the ProGait dataset for prosthesis-specific tasks. Our code is available at https://github.com/pittisl/ProGait and dataset at https://huggingface.co/datasets/ericyxy98/ProGait.

Federated AIoT uses distributed data on IoT devices to train AI models. However, in practical AIoT systems, heterogeneous devices cause data heterogeneity and varying amounts of device staleness, which can reduce model performance or increase federated training time. When addressing the impact of device delays, existing FL frameworks improperly consider it as independent from data heterogeneity. In this paper, we explore a scenario where device delays and data heterogeneity are closely correlated, and propose FedDC, a new technique to mitigate the impact of device delays in such cases. Our basic idea is to use gradient inversion to learn knowledge about device's local data distribution and use such knowledge to compensate the impact of device delays on devices' model updates. Experiment results on heterogeneous IoT devices show that FedDC can improve the FL performance by 34% with high amounts of device delays, without impairing the devices' local data privacy.

Multimodal reasoning by LLMs is critical to autonomous mobile systems, but the growing diversity of input data modalities prevents incorporating all modalities into LLMs. Instead, only the useful modalities should be adaptively involved at runtime, based on the current environmental contexts and task requirements. Existing work on runtime modality adaptation uses fixed connections between data encoders and LLM's input layer, but results in high training costs and ineffective cross-modal interaction. In this paper, we present MPnP, a new modality adaptation technique that connects data encoders to a flexible set of last LLM blocks and makes such latent connections fully trainable at runtime. Evaluation results show that MPnP has high compute and data efficiency, with 3.7× FLOPs reduction and 30% memory usage reduction compared to best baselines. It requires only few hundreds of training samples at runtime, and completes modality adaptation within few minutes on weak devices.

We propose FocusX, the first mobile-deployable system achieving artifact-free all-in-focus synthesis in dynamic scenes. Our approach introduces three key innovations: 1) For focal stack acquisition, our depth prior-based dynamic focusing method that adaptively selects focus distances using real-time scene depth distribution analysis and depth-of-field constrained spatial clustering, reducing redundant captures while ensuring full depth coverage; 2) To reduce pixel misalignment caused by lens breathing, we adopt a one-time offline calibration to map the relationship between field-of-view and focus distance, aligning the images by cropping accordingly; 3) We design the Diff-MotionAIFNet, a conditional diffusion-based model that decouples moving-static components for artifact-free AIF reconstruction in dynamic scene while preserving scene fidelity. We further contribute DynaAIFSet, containing 5,500 dynamic scenes (120K images) for training and evaluation. Experiments show FocusX achieves state-of-the-art performance, outperforming baselines up by 59.6% in SSIM and 49.1% in PSNR, respectively. The deployment latency of FocusX is 4.8s on Honor Magic7 Pro. This work bridges computational photography theory with mobile implementation constraints, delivering practical AIF enhancement for user-generated content.

Generative models have gained significant attention for their ability to produce realistic synthetic data that supplements the quantity of real-world datasets. While recent studies show performance improvements in wireless sensing tasks by incorporating all synthetic data into training sets, the quality of synthetic data remains unpredictable and the resulting performance gains are not guaranteed. To address this gap, we propose tractable and generalizable metrics to quantify quality attributes of synthetic data—affinity and diversity. Our assessment reveals prevalent affinity limitation in current wireless synthetic data, leading to mislabeled data and degraded task performance. We attribute the quality limitation to generative models' lack of awareness of untrained conditions and domain-specific processing. To mitigate these issues, we introduce SynCheck, a quality-guided synthetic data utilization scheme that refines synthetic data quality during task model training. Our evaluation demonstrates that SynCheck consistently outperforms quality-oblivious utilization of synthetic data, and achieves 4.3% performance improvement even when the previous utilization degrades performance by 13.4%.

Personalization of Large Language Models (LLMs) is important in practical applications to accommodate the individual needs of different mobile users. Due to data privacy concerns, LLM personalization often needs to be locally done at the user's mobile device, but such on-device personalization is constrained by both the limitation of on-device compute power and insufficiency of user's personal data. In this paper, we address these constraints by fine-tuning an already personalized LLM with user's personal data, and present XPerT, a new technique that ensure proper selection of such already personalized LLMs based on explainability about how they were being fine-tuned. We implemented and evaluated XPerT on various smartphone models with mainstream LLMs, and experiment results show that XPerT reduces the computation costs of on-device LLM personalization by 83%, and improves its data efficiency by 51%.

Hybrid memory systems exhibit disparities in their heterogeneous memory components' access speeds. Dynamic page scheduling to ensure memory access predominantly occurs in the faster memory components is essential for optimizing the performance of hybrid memory systems. Recent works attempt to optimize page scheduling by predicting their hotness using neural network models. However, they face two crucial challenges: the page explosion problem and the new pages problem. We propose an intelligent hybrid memory scheduler driven by page pattern recognition to address these two challenges. Experimental results demonstrate that our approach outperforms state-of-the-art intelligent schedulers regarding effectiveness and cost.

Fine-tuning is essential to adapting pre-trained large language models to downstream applications. With the increasing popularity of LLM-enabled applications, fine-tuning has been performed intensively worldwide, incurring a tremendous amount of computing costs that correspond to big carbon footprint and environmental impact. Mitigating such environmental impact directly correlates to reducing the fine-tuning FLOPs. Existing fine-tuning schemes focus on either saving memory or reducing the overhead of computing weight updates, but cannot achieve sufficient FLOPs reduction due to their ignorance of the training cost in backpropagation. To address this limitation, in this paper we present GreenTrainer, a new technique that minimizes the FLOPs of LLM fine-tuning via adaptive backpropagation, which adaptively selects the most appropriate set of LLM tensors for fine-tuning based on their importance and backpropagation cost in training. Experiment results show that GreenTrainer can save up to 64\% training FLOPs compared to full fine-tuning, without any noticeable accuracy loss. Compared to the existing schemes such as Prefix Tuning and LoRA, GreenTrainer can achieve up to 4\% improvement of model accuracy, with on-par FLOPs reduction.

Image super-resolution (SR) is widely used on mobile devices to enhance user experience. However, neural networks used for SR are computationally expensive, posing challenges for mobile devices with limited computing power. A viable solution is to use heterogeneous processors on mobile devices, especially the specialized hardware AI accelerators, for SR computations, but the reduced arithmetic precision on AI accelerators can lead to degraded perceptual quality in upscaled images. To address this limitation, in this paper we present SR For Your Eyes (FYE-SR), a novel image SR technique that enhances the perceptual quality of upscaled images when using heterogeneous processors for SR computations. FYE-SR strategically splits the SR model and dispatches different layers to heterogeneous processors, to meet the time constraint of SR computations while minimizing the impact of AI accelerators on image quality. Experiment results show that FYE-SR outperforms the best baselines, improving perceptual image quality by up to 2×, or reducing SR computing latency by up to 5.6× with on-par image quality.

On-device training is essential for neural networks (NNs) to continuously adapt to new online data, but can be time-consuming due to the device's limited computing power. To speed up on-device training, existing schemes select trainable NN portion offline or conduct unrecoverable selection at runtime, but the evolution of trainable NN portion is constrained and cannot adapt to the current need for training. Instead, runtime adaptation of on-device training should be fully elastic, i.e., every NN substructure can be freely removed from or added to the trainable NN portion at any time in training. In this paper, we present ElasticTrainer, a new technique that enforces such elasticity to achieve the required training speedup with the minimum NN accuracy loss. Experiment results show that ElasticTrainer achieves up to 3.5× more training speedup in wall-clock time and reduces energy consumption by 2×-3× more compared to the existing schemes, without noticeable accuracy loss.

Remote monitoring and evaluation of pulmonary diseases via tele-medicine are important to disease diagnosis and management, but current telemedicine solutions have limited capability of objectively examining the airway's internal physiological conditions that are crucial to pulmonary disease evaluation. Existing solutions based on smartphone sensing are also limited to externally monitoring breath rates, respiratory events, or lung function. In this paper, we present PTEase, a new system design that addresses these limitations and uses commodity smartphones to examine the airway's internal physiological conditions. PTEase uses active acoustic sensing to measure the internal changes of lower airway caliber, and then leverages machine learning to analyze the sensory data for pulmonary disease evaluation. We implemented PTEase as a smartphone app, and verified its measurement error in lab-controlled settings as <10%. Clinical studies further showed that PTEase reaches 75% accuracy on disease prediction and 11%-15% errors in estimating lung function indices. Given that such accuracy is comparable with that in clinical practice using spirometry, PTEase can be reliably used as an assistive telemedicine tool for disease evaluation and monitoring.

Graphics Processing Unit (GPU) on smartphones is an effective target for hardware attacks. In this paper, we present a new side channel attack on mobile GPUs of Android smartphones, allowing an unprivileged attacker to eavesdrop the user's credentials, such as login usernames and passwords, from their inputs through on-screen keyboard. Our attack targets on Qualcomm Adreno GPUs and investigate the amount of GPU overdraw when rendering the popups of user's key presses of inputs. Such GPU overdraw caused by each key press corresponds to unique variations of selected GPU performance counters, from which these key presses can be accurately inferred. Experiment results from practical use on multiple models of Android smartphones show that our attack can correctly infer more than 80% of user's credential inputs, but incur negligible amounts of computing overhead and network traffic on the victim device. To counter this attack, this paper suggests mitigations of access control on GPU performance counters, or applying obfuscations on the values of GPU performance counters.

Facial expressions are important indicators of user needs that can be used in many interactive computing applications to adapt the system behaviors and settings. Current computing approaches to recognizing human facial expressions, however, either rely on con-tinuous camera recordings that are energy consuming, or require custom sensing hardware that are expensive and difficult to use on commodity systems. In this paper, we present FaceListener, a new sensing system that recognizes human facial expressions by only using commodity headphones. The basic idea of FaceListener is to transform the commodity headphone into an acoustic sensing device, which captures the face skin deformations caused by fa-cial muscle movements with different facial expressions. To ensure the recognition accuracy, FaceListener leverages the knowledge distillation technique to learn the subtle correlation between face skin deformation and the acoustic signal changes. Experiment re-sults over multiple human beings demonstrate that FaceListener can accurately recognize more than 80% of different facial expressions. FaceListener is highly energy efficient, and can well adapt to different headphone models, host systems and user activities.

With the wide adoption of AI applications, there is a pressing need of enabling real-time neural network (NN) inference on small embedded devices, but deploying NNs and achieving high performance of NN inference on these small devices is challenging due to their extremely weak capabilities. Although NN partitioning and offloading can contribute to such deployment, they are incapable of minimizing the local costs at embedded devices. Instead, we suggest to address this challenge via agile NN offloading, which migrates the required computations in NN offloading from online inference to offline learning. In this paper, we present AgileNN, a new NN offloading technique that achieves real-time NN inference on weak embedded devices by leveraging eXplainable AI techniques, so as to explicitly enforce feature sparsity during the training phase and minimize the online computation and communication costs. Experiment results show that AgileNN's inference latency is >6X lower than the existing schemes, ensuring that sensory data on embedded devices can be timely consumed. It also reduces the local device's resource consumption by >8X, without impairing the inference accuracy.

New wireless physical-layer designs are the key to improving wireless network performance. Adopting these new designs, however, requires modifications on wireless hardware and is difficult on commodity devices. In this paper, we show that this hardware modification in many cases can be avoided by TransFi, a new software technique that enables custom wireless PHY functionality on commodity WiFi transmitters via fine-grained emulation. Our basic insight is that many custom wireless signals can be emulated by manipulating the MAC payloads of WiFi MIMO streams and mixing the transmitted signals from these streams on the air. To perform such emulation, TransFi considers the target signal as a mixture of QAM constellation points on the complex plane, and reversely computes the MAC payload of each MIMO stream from one selected QAM constellation point. We implemented TransFi on commodity WiFi devices to emulate three custom wireless PHYs with diverse characteristics. Experiment results show that TransFi's accuracy of emulation is >90% when transmitting emulated data payloads at 11.4 Mbps (46x faster than existing methods), and the decoding error at this data rate is <1% (10x lower than existing methods).

Interference could result in significant performance degradation in WiFi networks. Most existing solutions to interference cancellation require extra RF hardware, which is usually infeasible in many low-power wireless scenarios. In this paper, we present AiFi, a new interference cancellation technique that can be applied to commodity WiFi devices without using any extra RF hardware. The key idea of AiFi is to retrieve knowledge about interference from the locally available physical-layer (PHY) information at the WiFi receiver, including the pilot information (PI) and the channel state information (CSI). AiFi leverages the power of AI to address the possible ambiguity when estimating interference from these PHY information, and incorporates the domain knowledge about WiFi PHY to minimize the neural network complexity. Experiment results show that AiFi can correct 80% of bit errors due to interference and improves the MAC frame reception rate by 18x, with <1ms latency for interference cancellation in each frame.

Pulmonary diseases, such as asthma and Chronic Obstructive Pulmonary Disease (COPD), constitute a major public health challenge. The disease symptoms, including airway obstruction and inflammation, usually result in changes in airway mechanical properties, such as the caliber and impedance of the airway. To measure such airway properties for disease evaluation and diagnosis purposes, pulmonary function tests (PFT) has been widely adopted. However, most existing PFT systems require expensive and cumbersome hardware that are impossible to be used out of clinic. To allow out-clinic continuous pulmonary disease evaluation, in this paper we present AWARE, a new sensing and AI system that supports accurate and reliable PFT using commodity smartphones. AWARE uses a smartphone to transmit acoustic signals and reconstructs the profile of human airway based on the analysis of reflected acoustic waves captured from the smartphone's microphone. The subject's pulmonary condition is then evaluated by a multi-task learning model that integrates both the airway measurements and the subject's lung function records as the ground truth. Evaluations on 75 human subjects demonstrate that AWARE has the capability to achieve 80% accuracy on distinguishing between humans with healthy pulmonary function and with asthma symptoms.

Respiratory diseases have been a significant public health challenge. Efficient disease evaluation and monitoring call for daily spirometry tests, as an effective way of pulmonary function testing, out of clinic. This requirement, however, is hard to be satisfied due to the large size and high costs of current spirometry equipments. In this paper, we present SpiroSonic, a new system design that uses commodity smartphones to support complete, accurate yet reliable spirometry tests in regular home settings with various environmental and human factors. SpiroSonic measures the humans' chest wall motion via acoustic sensing and interprets such motion into lung function indices, based on the clinically validated correlation between them. We implemented SpiroSonic as a smartphone app, and verified SpiroSonic's monitoring error over healthy humans as <3%. Clinical studies further show that SpiroSonic reaches 5%-10% monitoring error among 83 pediatric patients. Given that the error of in-clinic spirometry is usually around 5%, SpiroSonic can be reliably used for disease tracking and evaluation out of clinic.

Stylus pens have been widely used with today's mobile devices to provide a convenient handwriting input method, but also bring a unique security vulnerability that may unveil the user's handwriting contents to a nearby eavesdropper. In this paper, we present MagHacker, a new sensing system that realizes such eavesdropping attack over commodity mobile devices, which monitor and analyze the magnetic field being produced by the stylus pen's internal magnet. MagHacker divides the continuous magnetometer readings into small segments that represent individual letters, and then translates these readings into writing trajectories for letter recognition. Experiment results over realistic handwritings from multiple human beings demonstrate that MagHacker can accurately eavesdrop more than 80% of handwriting with stylus pens, from a distance of 10cm. Only slight degradation in such accuracy is produced when the eavesdropping distance or the handwriting speed increases. MagHacker is highly energy efficient, and can well adapt to different stylus pen models and environmental contexts.

Performance of modern cognitive and interactive mobile applications highly depends on the transmission delay in the wireless link that is vital to supporting real-time wireless traffic. To eliminate wireless network congestion caused by large amounts of concurrent network traffic and minimize such transmission delay, traditional schemes adopt various flow control and QoS-aware traffic scheduling techniques, but fail when the amount of network traffic further increases. In this paper, we present a novel design of high-throughput wireless side channel, which operates concurrently with the existing wireless network channel over the same spectrum but dedicates to real-time traffic. Our key idea of realizing such a side channel is to exploit the excessive SNR margin in the wireless network to encode data as patterned interference. We design such patterned interference in the form of energy erasure over specific subcarriers in an OFDM-based wireless network, and achieve a data rate of 1.25 Mbps in the side channel without affecting the existing wireless network links. Experimental results over both software-defined radios and custom wireless hardware demonstrate the effectiveness of our side channel design in reducing the latency of real-time wireless traffic, while providing a sufficient data throughput for such traffic.

Recent diversification of mobile computing devices allows a mobile user to own multiple types of devices for different application scenarios, but also results in various restrictions on the performance and usability of these devices. A viable solution to such restriction is to incorporate and interconnect mobile devices towards a personal mobile cloud where these devices can complement each other via cooperative resource sharing, but is challenging due to the heterogeneity of mobile devices in both hardware and software aspects. In this paper, we propose a novel design of resource sharing framework to address these challenges and generically interconnect heterogeneous mobile devices. Our basic idea is to mask the hardware and software heterogeneity in mobile systems by exploiting the existing mobile OS services as the interface of resource sharing, and further develop the resource sharing framework as a middleware in the mobile OS. We have implemented our design over various mobile platforms with diverse characteristics and resource limits, and demonstrated that our design can efficiently support generic resource sharing among heterogeneous mobile devices without incurring significant system overhead or requiring individual system modification.

Many IoT applications have stringent requirements on wireless transmission delay, but have to compete for channel access with other wireless traffic. Traditional techniques enable multiple access to wireless channels, but yield severe delay when the channel is congested. In this paper, we present EasyPass, a wireless PHY technique that allows multiple IoT devices to simultaneously transmit data over a congested wireless link without being delayed. The key idea of EasyPass is to exploit the excessive SNR margin in a wireless channel as a dedicated side channel for IoT traffic, and allow multiple access to the side channel by separating signals from different transmitters on the air. We implemented EasyPass on software-defined radio platforms. Experiment results demonstrate that EasyPass reduces the data transmission delay in congested IoT networks by 90%, but provides a throughput up to 2.5 Mbps over a narrowband 20MHz wireless link that can be accessed by more than 100 IoT devices.

Cross-technology coexistence is crucial to avoid collisions of wireless transmissions and improve the efficiency of spectrum utilization in today's large-scale wireless network systems, especially the Internet of Things. However, existing approaches to cross-technology coexistence incur additional transmission delay and signal processing overhead, which are unaffordable by extremely weak wireless devices such as embedded sensors and computational RFIDs. These schemes hence fail when being applied to emerging application scenarios, such as smart cities and connected healthcare where weak devices play important roles. In this paper, we design and implement EmBee, a new wireless PHY technique that enables cross-technology coexistence at zero cost or performance loss to these extremely weak wireless devices. The basic idea of EmBee is to exploit the diversity of different wireless technologies' spectrum utilization, so as to adaptively reserve occupied spectrum from the strong devices for weak wireless devices' concurrent data transmissions. We have implemented EmBee over custom wireless hardware and evaluated EmBee under different wireless scenarios. Experiment results show that EmBee can effectively support ZigBee transmissions over a fully occupied WiFi channel without causing any extra delay, while only resulting in 10% WiFi throughput loss.

Virtual Reality (VR) improves the user's experience when interacting with the virtual world, and could revolutionarily transform the designs of many interactive systems. However, providing VR from untethered mobile devices is difficult due to their limited local capabilities. Existing VR solutions address this difficulty by rendering VR frames at remote computing facilities, but are limited to rendering every VR frame separately. A tremendous amount of VR frame data, hence, needs to be transmitted to mobile devices over low-bandwidth wireless links and seriously impairs VR performance. In this paper, we aim to remove this performance constraint on highly dynamic VR applications with complicated scenes and intensive user movement, by adaptively reusing the redundant VR pixels across multiple VR frames. We leverage the unique characteristics of image warping used in current VR applications, and fundamentally expand the scope of image warping to the entire VR lifespan to precisely capture the fluctuations of VR scene due to VR dynamics. We implemented our design over Android OS and Unity VR application engine, and demonstrated that our design can maximize the mobile VR performance over highly dynamic VR scenarios with 95% less amount of VR frame data being transmitted, by completely removing the pixel redundancy across VR frames.

Device-free acoustic motion tracking allows a commodity mobile device to precisely track the human user's motion, without applying any extra hardware tracker on the human body. Most of current device-free acoustic motion tracking systems, however, are limited to tracking the motion of small parts of the human body with negligible sizes, such as human fingers. Their accuracy of motion tracking will significantly degrade when being applied to targets with large sizes, such as humans' hands, arms or body trunk. We envision the key reason to such degradation as the target size's significant impact on the pattern of the reflected acoustic signal, and develop analytical modeling of such reflected acoustic signal from large targets. Based on such modeling, we present a new system called Acoustic Tracking over targets with LArge Sizes (ATLAS), which ensures precise motion tracking over large targets by correctly interpreting the reflected acoustic signal and extracting the phase from the signal. Experiment results over commodity Android smartphones show that ATLAS can reduce the error of motion tracking by more than 75%, when being applied to targets with heterogeneous sizes in practice.

Mobile devices differ from desktop computers in that they have a limited power source, a battery, and they tend to spend more CPU time on the graphical user interface (GUI). These two facts force us to consider different software approaches in the mobile device kernel that can conserve battery life and reduce latency, which is the duration of time between the inception of an event and the reaction to the event. One area to consider is a software package called the display server. The display server is middleware that handles all GUI activities between an application and the operating system, such as event handling and drawing to the screen. In both desktop and mobile devices, the display server is located in the application layer. However, the kernel layer contains most of the information needed for handling events and drawing graphics, which forces the application-level display server to make a series of system calls in order to coordinate events and to draw graphics. These calls interrupt the CPU which can increase both latency and power consumption, and also require the kernel to maintain event queues that duplicate event queues in the display server. A further drawback of placing the display server in the application layer is that the display server contains most of the information required to efficiently schedule the application and this information is not communicated to existing kernels, meaning that GUI-oriented applications are scheduled less efficiently than they might be, which further increases power consumption. We propose moving the display server to the kernel layer, so that it has direct access to many of the event queues and hardware rendering systems without having to interrupt the CPU. This adjustment has allowed us to implement two power saving strategies, discussed in other papers, that streamline the event system and improve the scheduler. The combination of these two techniques reduces power consumption by an average of 30 percent and latency by an average of 17 ms. Even without the implementation of these power saving techniques, the KDS increases battery life by 4.35 percent or on average about 10 extra minutes for a typical mobile phone or 30 extra minutes for a typical tablet computer. It also reduces latency by 1.1 milliseconds.

Internet of Things has stringent requirements on the wireless network throughput for timely transmission of the big data being produced. In order to maximize the throughput over dynamic fluctuations of wireless channel quality, current wireless systems adapt the link rate to the instantaneous channel condition, but fail to fully utilize the channel capacity due to the discrete choices of available link rates and the gap between these rates. Instead, in this paper we present vMod, a lightweight and practical solution towards maximum wireless network throughput by redesigning the wireless link rates from discrete to continuous. The key idea of vMod is to modulate a fractional number of data bits into each symbol by employing the Variable-Length Code (VLC), which is able to statistically yield any link rate. We implemented vMod on software-defined radio platforms. Experiment results demonstrate that under highly dynamic wireless network conditions, vMod greatly improves the WiFi throughput by 30% over a single narrowband link, but incurs only negligible overhead.

Recent diversification of mobile computing devices allows a mobile user to own multiple types of devices for different application scenarios, but also results in various restrictions on the performance and usability of these devices. A viable solution to such restriction is to incorporate and interconnect mobile devices towards a personal mobile cloud where these devices can complement each other via cooperative resource sharing, but is challenging due to the heterogeneity of mobile devices in both hardware and software aspects. In this paper, we propose a novel design of resource sharing framework to address these challenges and generically interconnect heterogeneous mobile devices. Our basic idea is to mask the hardware and software heterogeneity in mobile systems by exploiting the existing mobile OS services as the interface of resource sharing, and further develop the resource sharing framework as a middleware in the mobile OS. We have implemented our design over various mobile platforms with diverse characteristics and resource limits, and demonstrated that our design can efficiently support generic resource sharing among heterogeneous mobile devices without incurring significant system overhead or requiring individual system modification.

Mobile Cloud Computing (MCC) is of particular importance to address the conflict between the increasing complexity of user applications and the limited lifespan of mobile device's battery, by offloading the computational workloads from local devices to the remote cloud. Current offloading schemes either require the programmer's annotations, which restricts its wide application; or transmits too much unnecessary data, resulting bandwidth, and energy waste. In this paper, we propose a novel method-level offloading methodology to offload local computational workload with as least data transmission as possible. Our basic idea is to identify the contexts which are necessary to the method execution by parsing application binaries in advance and applying this parsing result to selectively migrate heap data while allowing successful method execution remotely. To further improve the efficiency of such offline parsing of application binaries, our scheme also conducts one-time parsing to all the mobile OS libraries and reuses these parsing results for different user applications. We have implemented our design over the Dalvik Virtual Machine of Android OS. Our experiments and evaluation against applications downloaded from Google Play show that our approach can save data transmission significantly comparing to existing schemes.

Scheduling in wireless networks is critical to maximize the network throughput by avoiding interference among wireless links, and is usually formulated as solving the NP-hard Maximum Weighted Independent Set (MWIS) problem over a network confiict graph. Existing scheduling algorithms are designed to provide approximations to global optimality via distributed operations in wireless networks, but will frequently reschedule the entire network in cases of network dynamics regardless of the actual network area being affected by these dynamics. Such repetitive rescheduling results in a large amount of computation and communication overhead, most of which may be, however, unnecessarily incurred over the wireless links that remain unchanged. To reduce such overhead and improve the scheduling cost-effectiveness, in this paper we develop distributed algorithms that adaptively constrain network scheduling within the limited scope where network dynamics occur. The scheduling results from such limited operations are then combined with the previous scheduling results over the remaining portions of the network, hence still providing guaranteed network throughput. The performance of our proposed algorithms has been validated by formal analysis, and been verified by both numerical studies and real-world experiments.

The performance of mobile computing would be significantly improved by leveraging cloud computing and migrating mobile workloads for remote execution at the cloud. In this paper, to efficiently handle the peak load and satisfy the requirements of remote program execution, we propose to deploy cloud servers at the network edge and design the edge cloud as a tree hierarchy of geo-distributed servers, so as to efficiently utilize the cloud resources to serve the peak loads from mobile users. The hierarchical architecture of edge cloud enables aggregation of the peak loads across different tiers of cloud servers to maximize the amount of mobile workloads being served. To ensure efficient utilization of cloud resources, we further propose a workload placement algorithm that decides which edge cloud servers mobile programs are placed on and how much computational capacity is provisioned to execute each program. The performance of our proposed hierarchical edge cloud architecture on serving mobile workloads is evaluated by formal analysis, small-scale system experimentation, and large-scale trace-based simulations.

Traditional power grid is not resistant to severe weather conditions, especially in remote areas. For some areas with few people, such as islands, it is difficult and expensive to maintain their connectivity to the traditional power grid. Therefore, a self-sustainable microgrid is desired. However, given the limited local energy storage and energy generation, it is extremely challenging for a microgrid to balance the power demand and generation in real-time. To realize the real-time power quality monitoring, the power quality information of microgrid, such as voltage, frequency and phase angle in each home, needs to be collected in real- time. Furthermore, the unreliable sensing results and data collection in a microgrid make the real-time data collection more difficult. To address these challenges, we designed an accurate real-time power quality data sensing hardware to sense the voltage, frequency and phase angle in each home. A novel data management technique is also proposed to reconstruct the missing data caused by unreliable sensing. We implemented our system over off-the-shelf smartphones with a few peripheral hardware components, and realized an accuracy of 1.7 mHz and 0.01 rad for frequency and phase angle monitoring, respectively. We also show our data management technique can reconstruct the missing data with more than 99% accuracy.

Performance of modern cognitive and interactive mobile applications highly depends on the data transmission delay in the wireless link that is vital to supporting real-time wireless traffic. To eliminate wireless network congestion caused by large amounts of concurrent network traffic and support such real-time traffic, traditional schemes adopt various flow control and QoS-aware traffic scheduling techniques, but fail when the amount of network traffic further increases. In this paper, we present a novel design of high-throughput wireless side channel, which operates concurrently with the existing wireless network channel over the same spectrum but dedicates to real-time traffic. Our key idea of realizing such a side channel is to exploit the excessive SNR margin in the wireless network to encode data as patterned interference. We design such patterned interference in form of energy erasure over specific subcarriers in an OFDM-based wireless network, and achieve a data rate of 1.25 Mbps in the side channel without affecting the existing wireless network links. Experimental results over software-defined radio platforms demonstrate the effectiveness of our side channel design in reducing the latency of real-time wireless traffic, while providing sufficient data throughput for such traffic.

Wireless rechargeable sensor networks have recently emerged as a promising platform that can effectively solve the power constraint problem suffered by traditional battery powered systems. The problem of determining the best charging routes for maximizing charging efficiency has been studied extensively. However, the task assignment problem, which plays a crucial role in efficiently utilizing the harvested energy and thus minimize the charging delay, has received rather limited attention. In this paper, we study the problem of assigning a given set of tasks in a wireless rechargeable sensor network while maximizing the charger's velocity to minimize the charging delay. We first propose an online task assignment algorithm, namely Lower Bound assignment (LB), that yields a quantifiable lower bound on the charging velocity while guaranteeing a feasible assignment. This algorithm further enables the transformation of our considered task assignment problem into a variation of the classical multiple knapsack problem. We then present a fully polynomial-time approximation scheme with a (2+ε)-approximation ratio, namely ACT, that is built upon an existing greedy algorithm designed for the original knapsack problem. Extensive experimental results presented herein demonstrate that ACT is able to achieve near-optimal performance in most cases, and can achieve more than 15% performance improvement compared to the baseline algorithms.

Nowadays, mobile phones are usually embedded with powerful cameras. Due to the convenience of carrying mobile phones, an increasing number of people use mobile phones to take photos anytime and anywhere. However, when a user takes a photo of a scenery, a building or a target person, sometimes an unexpected stranger is also included in the photo. Such photos reveal where the stranger has been and thus can breach his privacy. This problem has received little attention in the literature. In this paper, we propose PrivacyCamera, a cooperative system to protect the stranger's privacy in the above scenario. Through cooperation between the photographer and the stranger, the system can automatically blur the stranger's face in the photo upon the stranger's request when the photo is being taken. This paper describes the design, analysis, prototype implementation, and experimental evaluation of the system. Experiments show that PrivacyCamera can effectively protect stranger's privacy in an efficient way.

In cognitive radio networks, unlicensed users can use under-utilized licensed spectrum to achieve substantial performance improvement. To avoid interference with licensed users, unlicensed users must vacate the spectrum when it is accessed by licensed (primary) users. Since it takes some time for unlicensed users to switch to other available channels, the ongoing data transmissions may have to be interrupted and the transmission delay can be significantly increased. This makes it hard for cognitive radio networks to meet the delay constraints of many applications. In this paper, we develop caching techniques to address this problem. We formulate the cache placement problem in cognitive radio networks as an optimization problem, where the goal is to minimize the total cost, subject to some delay constraint, i.e., the data access delay can be statistically bounded. To solve this problem, we propose a cost-based approach to minimize the caching cost, and design a delay-based approach to satisfy the delay constraint. Then, we combine them and propose a distributed hybrid approach to minimize the caching cost subject to the delay constraint. Simulation results show that our approaches outperform existing caching solutions in terms of total cost and delay constraint, and the hybrid approach performs the best among the approaches satisfying the delay constraint.

The recent popularization of hand-held mobile devices, such as smartphones, enables the inter-connectivity among mobile users without the support of Internet infrastructure. When mobile users move and contact each other opportunistically, they form a delay tolerant network (DTN), which can be exploited to share data among them. Data replication is one of the common techniques for such data sharing. However, the unstable network topology and limited contact duration in DTNs make it difficult to directly apply traditional data replication schemes. In this paper, we recognize the deficiency of existing data replication schemes which treat the complete data item as the replication unit, and propose to replicate data at the packet level using erasure coding techniques. Our study consists of two cases based on the operating spectrum: unlicensed spectrum and licensed spectrum. For both cases, we analytically formulate the data replication problem as a mixed integer programming problem and propose a practical algorithm which operates in a fully distributed manner. Extensive simulations on both synthetic and realistic traces show that our scheme outperforms other existing replication schemes in terms of successful data retrieval probability in various scenarios.

Recent developments in embedded hardware have empowered human experiences through pervasive computing. While embedded systems are becoming more powerful, they still fall short when faced with users' growing desire for running more resource-demanding applications. To bridge this gap, one solution is to leverage powerful resources residing at remote sites by performing computation offloading. Unfortunately, the state-of-the-art offloading frameworks cannot be applied in many embedded systems supporting applications with soft real-time (SRT) constraints or high delay sensitivity, as they typically optimize response times on a "best-effort" basis using heuristics. This paper establishes a soft real-time offloading framework that optimizes the resource utilization of the embedded system while analytically guaranteeing SRT schedulability. The key idea behind the proposed framework is to view offloading-induced delays as suspensions occurring at the local embedded system side, which allows a task being offloaded to be modelled as a suspending task and thus existing SRT suspension-aware scheduling and analysis techniques to be leveraged. Based on this idea, we propose an offloading algorithm, namely Real-time Offloading Decision-making Algorithm (RODA), to make offloading decisions such that SRT schedulability of the task system can be ensured. The optimality properties of RODA have been proved on both uniprocessors and multiprocessors. We conducted extensive simulations on evaluating schedulability and implemented a case study offloading system on top of real hardware to test runtime response time performance. Results demonstrated that RODA is superior to existing performance-driven offloading algorithms, particularly under heavy workloads.

Computationally intensive applications may quickly drain mobile device batteries. One viable solution to address this problem utilizes computation offloading. The tradeoff is that computation offloading introduces additional communication, with a corresponding energy cost. Yet, previous research into computation offloading has failed to account for the special characteristics of cellular networks that impact mobile device energy consumption. In this paper, we aim to develop energy efficient computation offloading algorithms for cellular networks. We analyze the effects of the long tail problem on task offloading, formalize the computation offloading problem, and use Dijkstra's algorithm to find the optimal decision. Since this optimal solution relies on perfect knowledge of future tasks, we further propose an online algorithm for offloading. We have implemented this latter algorithm on Android-based smartphones. Both experimental results from this implementation and trace-driven simulation show that our algorithm can significantly reduce the energy of computation offloading in cellular networks.

Mobile Cloud Computing (MCC) is of particular importance to address the contradiction between the increasing complexity of user applications and the limited lifespan of mobile device's battery, by offloading the computational workloads from local devices to the remote cloud. Current offloading schemes either require the programmer's annotations, which restricts its wide application; or transmits too much unnecessary data, resulting bandwidth and energy waste. In this paper, we propose a novel method-level offloading methodology to offload local computational workload with as least data transmission as possible. Our basic idea is to identify the contexts which are necessary to the method execution by parsing application binaries in advance and applying this parsing result to selectively migrate heap data while allowing successful method execution remotely. Our implementation of this design is built upon Dalvik Virtual Machine. Our experiments and evaluation against applications downloaded from Google Play show that our approach can save data transmission significantly comparing to existing schemes.

Opportunistic mobile networks consist of mobile devices which are intermittently connected via short-range radios. Forwarding in such networks relies on selecting relays to carry and deliver data to destinations upon opportunistic contacts. Due to the intermittent network connectivity, relays in current forwarding schemes are selected separately in a distributed manner. The contact capabilities of relays hence may overlap when they contact the same nodes and cause forwarding redundancy. This redundancy reduces the efficiency of resource utilization in the network, and may impair the forwarding performance if being unconsciously ignored. In this paper, based on investigation results on the characteristics of forwarding redundancy in realistic mobile networks, we propose methods to eliminate unnecessary forwarding redundancy and ensure efficient utilization of network resources. We first develop techniques to eliminate forwarding redundancy with global network information, and then improve these techniques to be operable in a fully distributed manner with limited network information. We furthermore propose adaptive forwarding strategy to intentionally control the amount of forwarding redundancy and satisfy the required forwarding performance with minimum cost. Extensive trace-driven evaluations show that our schemes effectively enhance forwarding performance with much lower cost.

Anti-virus systems developed by different vendors often demonstrate strong discrepancies in how they name malware, which signficantly hinders malware information sharing. While existing work has proposed a plethora of malware naming standards, most anti-virus vendors were reluctant to change their own naming conventions. In this paper we explore a new, more pragmatic alternative. We propose to exploit the correlation between malware naming of different anti-virus systems to create their consensus classification, through which these systems can share malware information without modifying their naming conventions. Specifically we present Latin, a novel classification integration framework leveraging the correspondence between participating anti-virus systems as reflected in heterogeneous information sources at instance-instance, instance-name, and name-name levels. We provide results from extensive experimental studies using real malware datasets and concrete use cases to verify the efficacy of Latin in supporting cross-system malware information sharing.

Mobile Cloud Computing (MCC) bridges the gap between limited capabilities of mobile devices and the increasing users' demand of mobile multimedia applications, by offloading the computational workloads from local devices to the remote cloud. Current MCC research focuses on making offloading decisions over different methods of a MCC application, but may inappropriately increase the energy consumption if having transmitted a large amount of program states over expensive wireless channels. Limited research has been done on avoiding such energy waste by exploiting the dynamic patterns of applications' run-time execution for workload offloading. In this paper, we adaptively offload the local computational workload with respect to the run-time application dynamics. Our basic idea is to formulate the dynamic executions of user applications using a semi-Markov model, and to further make offloading decisions based on probabilistic estimations of the offloading operation's energy saving. Such estimation is motivated by experimental investigations over practical smart phone applications, and then builds on analytical modeling of methods' execution times and offloading expenses. Systematic evaluations show that our scheme significantly improves the efficiency of workload offloading compared to existing schemes over various smart phone applications.

Opportunistic mobile networks consist of mobile devices which are intermittently connected via short-range radios. Forwarding in such networks relies on selecting relays to carry and deliver data to destinations upon opportunistic contacts. Due to the intermittent network connectivity, relays in current forwarding schemes are selected separately in a distributed manner. The contact capabilities of relays hence may overlap when they contact the same nodes and cause forwarding redundancy. This redundancy reduces the efficiency of resource utilization in the network, and may impair the forwarding performance if being ignored. In this paper, based on experimental investigations on the characteristics of forwarding redundancy in realistic mobile networks, we propose methods to eliminate unnecessary forwarding redundancy and ensure efficient utilization of network resources. We first develop techniques to eliminate forwarding redundancy with global network information, and then improve these techniques to be operable in a fully distributed manner with limited network information.

In cognitive radio networks, unlicensed users can use under-utilized licensed spectrum to achieve substantial performance improvement. To avoid interference with licensed users, unlicensed users must vacate the spectrum when it is accessed by licensed (primary) users. Since it takes some time for unlicensed users to switch to other available channels, the ongoing data transmissions may have to be interrupted and the transmission delay can be significantly increased. This makes it hard for cognitive radio networks to meet the delay constraints of many applications. To the best of our knowledge, we are the first to use caching techniques to address this problem. We formulate the cache placement problem in cognitive radio networks as an optimization problem, where the goal is to minimize the total cost, subject to some delay constraint, i.e., the data access delay can be statistically bounded. To solve this problem, we propose three approaches: cost-based, delay-based, and hybrid. Simulation results show that our approaches outperform existing caching solutions in terms of total cost and delay constraint, and the hybrid approach performs the best among the approaches satisfying the delay constraint.

In this demo, we present MoodMagician, a pervasive and unobtrusive mobile phone system for inferring human emotions through the recording, processing, and analysis of the real-time streaming Galvanic Skin Response (GSR) signal from human bodies. Being different from traditional multimodal emotion sensing systems which rely on data from multiple sensing sources and may hence interfere with people's daily life, our proposed system is able to detect various categories of human emotions using single GSR signal, which is captured by compact and wearable mobile sensing devices in an unobtrusive fashion. The proposed system has been evaluated by well-designed practical experiments to recognize human emotions. The recognition accuracy of each emotion can be up to 70% through the development of effective preprocessing algorithms and the extraction of representative features from the GSR signals.

Disruption tolerant networks (DTNs) are characterized by low node density, unpredictable node mobility, and lack of global network information. Most of current research efforts in DTNs focus on data forwarding, but only limited work has been done on providing efficient data access to mobile users. In this paper, we propose a novel approach to support cooperative caching in DTNs, which enables the sharing and coordination of cached data among multiple nodes and reduces data access delay. Our basic idea is to intentionally cache data at a set of network central locations (NCLs), which can be easily accessed by other nodes in the network. We propose an efficient scheme that ensures appropriate NCL selection based on a probabilistic selection metric and coordinates multiple caching nodes to optimize the tradeoff between data accessibility and caching overhead. Extensive trace-driven simulations show that our approach significantly improves data access performance compared to existing schemes.

Cellular networks (e.g., 3G) are currently facing severe traffic overload problems caused by excessive traffic demands. Offloading part of the cellular traffic through other forms of networks, such as Delay Tolerant Networks (DTNs) and WiFi hotspots, is a promising solution. However, since these networks can only provide intermittent connectivity to mobile users, utilizing them for cellular traffic offloading may result in a nonnegligible delay. As the delay increases, the users' satisfaction decreases. In this paper, we investigate the tradeoff between the amount of traffic being offloaded and the users' satisfaction. We provide a novel incentive framework to motivate users to leverage their delay tolerance for cellular traffic offloading. To minimize the incentive cost given an offloading target, users with high delay tolerance and large offloading potential should be prioritized for traffic offloading. To effectively capture the dynamic characteristics of users' delay tolerance, our incentive framework is based on reverse auction to let users proactively express their delay tolerance by submitting bids. We further illustrate how to predict the offloading potential of the users by using stochastic analysis for both DTN and WiFi cases. Extensive trace-driven simulations verify the efficiency of our incentive framework for cellular traffic offloading.

Opportunistic mobile networks consist of mobile devices which only communicate when they opportunistically contact each other. Periodic contact probing is required to facilitate opportunistic communication, but seriously reduces the limited battery life of mobile devices. Current research efforts on reducing energy consumption of contact probing are restricted to optimize the probing interval, but are insufficient for energy-efficient opportunistic communication. In this paper, we propose novel techniques to adaptively schedule wakeup periods of mobile nodes between their inter-contact times. A node stays asleep during inter-contact times when contact probing is unnecessary, and only wakes up when a contact with another node is likely to happen. Our approach probabilistically predicts node contacts in the future, and analytically balances between energy consumption for contact probing and performance of opportunistic communication. Extensive trace-driven simulations show that our approach significantly improves energy efficiency of opportunistic communication compared to existing schemes.

Disruption Tolerant Networks (DTNs) utilize the mobility of nodes and the opportunistic contacts among nodes for data communications. Due to the limitation in network resources such as contact opportunity and buffer space, DTNs are vulnerable to flood attacks in which attackers send as many packets or packet replicas as possible to the network, in order to deplete or overuse the limited network resources. In this paper, we employ rate limiting to defend against flood attacks in DTNs, such that each node has a limit over the number of packets that it can generate in each time interval and a limit over the number of replicas that it can generate for each packet. We propose a distributed scheme to detect if a node has violated its rate limits. To address the challenge that it is difficult to count all the packets or replicas sent by a node due to lack of communication infrastructure, our detection adopts claim-carry-and-check: each node itself counts the number of packets or replicas that it has sent and claims the count to other nodes; the receiving nodes carry the claims when they move, and cross-check if their carried claims are inconsistent when they contact. The claim structure uses the pigeonhole principle to guarantee that an attacker will make inconsistent claims which may lead to detection. We provide rigorous analysis on the probability of detection, and evaluate the effectiveness and efficiency of our scheme with extensive trace-driven simulations.

Unpredictable node mobility, low node density, and lack of global information make it challenging to achieve effective data forwarding in Delay-Tolerant Networks (DTNs). Most of the current data forwarding schemes choose the nodes with the best cumulative capability of contacting others as relays to carry and forward data, but these nodes may not be the best relay choices within a short time period due to the heterogeneity of transient node contact characteristics. In this paper, we propose a novel approach to improve the performance of data forwarding with a short time constraint in DTNs by exploiting the transient social contact patterns. These patterns represent the transient characteristics of contact distribution, network connectivity and social community structure in DTNs, and we provide analytical formulations on these patterns based on experimental studies of realistic DTN traces. We then propose appropriate forwarding metrics based on these patterns to improve the effectiveness of data forwarding. When applied to various data forwarding strategies, our proposed forwarding metrics achieve much better performance compared to existing schemes with similar forwarding cost.

In this paper, we consider the issue of data broadcasting in mobile social networks (MSNets). The objective is to broadcast data from a superuser to other users in the network. There are two main challenges under this paradigm, namely 1) how to represent and characterize user mobility in realistic MSNets; 2) given the knowledge of regular users' movements, how to design an efficient superuser route to broadcast data actively. We first explore several realistic data sets to reveal both geographic and social regularities of human mobility, and further propose the concepts of geocommunity and geocentrality into MSNet analysis. Then, we employ a semi-Markov process to model user mobility based on the geocommunity structure of the network. Correspondingly, the geocentrality indicating the “dynamic user density” of each geocommunity can be derived from the semi-Markov model. Finally, considering the geocentrality information, we provide different route algorithms to cater to the superuser that wants to either minimize total duration or maximize dissemination ratio. To the best of our knowledge, this work is the first to study data broadcasting in a realistic MSNet setting. Extensive trace-driven simulations show that our approach consistently outperforms other existing superuser route design algorithms in terms of dissemination ratio and energy efficiency.

Opportunistic mobile networks consist of personal mobile devices which are intermittently connected with each other. Data access can be provided to these devices via cooperative caching without support from the cellular network infrastructure, but only limited research has been done on maintaining the freshness of cached data which may be refreshed periodically and is subject to expiration. In this paper, we propose a scheme to efficiently maintain cache freshness. Our basic idea is to let each caching node be only responsible for refreshing a specific set of caching nodes, so as to maintain cache freshness in a distributed and hierarchical manner. Probabilistic replication methods are also proposed to analytically ensure that the freshness requirements of cached data are satisfied. Extensive trace driven simulations show that our scheme significantly improves cache freshness, and hence ensures the validity of data access provided to mobile users.

We consider the problem of diffusing cached content in an intermittently connected mobile network, starting from a given initial configuration to a desirable goal state where all nodes interested in particular contents have a copy of their desired contents. The goal is to minimize the time taken for the diffusion process to terminate at a goal state. Due to bandwidth and storage constraints, whenever two nodes encounter each other, they must decide which content if any to transfer to each other. While most prior work on this topic has focused on practically realizable heuristics for this problem, we take a more formal approach. Our main contribution is to show that, assuming global state information is available, this problem can be formulated as a stochastic shortest path problem, which is a kind of Markov decision process (MDP). Using this formulation, we numerically explore some small-scale examples for which we are able to obtain the optimal solution. The results show that the optimal diffusion strategy is very much a function of the underlying encounter graph.

Node mobility and end-to-end disconnections in disruption-tolerant networks (DTNs) greatly impair the effectiveness of data forwarding. Although social-based approaches can address the problem, most existing solutions only focus on forwarding data to a single destination. In this paper, we study multicast with single and multiple data items in DTNs from a social network perspective, develop analytical models for multicast relay selection, and furthermore investigate the essential difference between multicast and unicast in DTNs. The proposed approach selects relays according to their capabilities, measured by social-based metrics, for forwarding data to the destinations. The design of social-based metrics exploits social network concepts such as node centrality and social community, and the selected relays ensure achieving the required data delivery ratio within the given time constraint. Extensive trace-driven simulations show that the proposed approach has similar data delivery ratio and delay to that of Epidemic routing, but significantly reduces data forwarding cost, measured by the number of relays used.

Disruption Tolerant Networks (DTNs) are characterized by the low node density, unpredictable node mobility and lack of global network information. Most of current research efforts in DTNs focus on data forwarding, but only limited work has been done on providing effective data access to mobile users. In this paper, we propose a novel approach to support cooperative caching in DTNs, which enables the sharing and coordination of cached data among multiple nodes and reduces data access delay. Our basic idea is to intentionally cache data at a set of Network Central Locations (NCLs), which can be easily accessed by other nodes in the network. We propose an effective scheme which ensures appropriate NCL selection based on a probabilistic selection metric, and coordinate multiple caching nodes to optimize trade off between data accessibility and caching overhead. Extensive trace-driven simulations show that our scheme significantly improves data access performance compared to existing schemes.

3G networks are currently facing severe traffic overload problems caused by excessive demands of mobile users. Offloading part of the 3G traffic through other forms of networks, such as Delay Tolerant Networks (DTNs), WiFi hotspots, and Femtocells, is a promising solution. However, since these networks can only provide intermittent and opportunistic connectivity to mobile users, utilizing them for 3G traffic offloading may result in a non-negligible delay. As the delay increases, the users' satisfaction decreases. In this paper, we investigate the tradeoff between the amount of traffic being offloaded and the users' satisfaction. We provide a novel incentive framework to motivate users to leverage their delay tolerance for 3G traffic offloading. To minimize the incentive cost given an offloading target, users with high delay tolerance and large offloading potential should be prioritized for traffic offloading. To effectively capture the dynamic characteristics of users' delay tolerance, our incentive framework is based on reverse auction to let users proactively express their delay tolerance by submitting bids. We further take DTN as a case study to illustrate how to predict the offloading potential of the users by using stochastic analysis. Extensive trace-driven simulations verify the efficiency of our incentive framework for 3G traffic offloading.

The recent popularization of hand-held mobile devices, such as smartphones, enables the inter-connectivity among mobile users without the support of Internet infrastructure. When mobile users move and contact each other opportunistically, they form a Delay Tolerant Network (DTN), which can be exploited to share data among them. Data replication is one of the common techniques for such data sharing. However, the unstable network topology and limited contact duration in DTNs make it difficult to directly apply traditional data replication schemes. Although there are a few existing studies on data replication in DTNs, they generally ignore the contact duration limits. In this paper, we recognize the deficiency of existing data replication schemes which treat the complete data item as the replication unit, and propose to replicate data at the packet level. We analytically formulate the contact duration aware data replication problem and give a centralized solution to better utilize the limited storage buffers and the contact opportunities. We further propose a practical contact Duration Aware Replication Algorithm (DARA) which operates in a fully distributed manner and reduces the computational complexity. Extensive simulations on both synthetic and realistic traces show that our distributed scheme achieves close-to-optimal performance, and outperforms other existing replication schemes.

Data dissemination is useful for many applications of Disruption Tolerant Networks (DTNs). Current data dissemination schemes are generally network-centric ignoring user interests. In this paper, we propose a novel approach for user-centric data dissemination in DTNs, which considers satisfying user interests and maximizes the cost-effectiveness of data dissemination. Our approach is based on a social centrality metric, which considers the social contact patterns and interests of mobile users simultaneously, and thus ensures effective relay selection. The performance of our approach is evaluated from both theoretical and experimental perspectives. By formal analysis, we show the lower bound on the cost-effectiveness of data dissemination, and analytically investigate the tradeoff between the effectiveness of relay selection and the overhead of maintaining network information. By trace-driven simulations, we show that our approach achieves better cost-effectiveness than existing data dissemination schemes.

Effective data forwarding in Delay Tolerant Networks (DTNs) is challenging, due to the low node density, unpredictable node mobility and lack of global information in such networks. Most of the current data forwarding schemes choose the nodes with the best cumulative capability of contacting others as relays to carry and forward data, but these nodes may not be the best relay choices within a short time period, due to the heterogeneity of the transient node contact patterns. In this paper, we propose a novel approach to improve the performance of data forwarding in DTNs by exploiting the transient node contact patterns. We formulate the transient node contact patterns based on experimental studies of realistic DTN traces, and propose appropriate forwarding metrics based on these patterns to improve the effectiveness of data forwarding decision. When applied to various data forwarding strategies, our proposed forwarding metrics achieve much better performance compared to existing schemes with similar forwarding cost.

In mobile social networks (MSNets), data dissemination is an important topic, which has not been widely investigated yet. Active data dissemination is a networking paradigm where a superuser intentionally facilitates the connectivity in the network. One of the key challenges under this paradigm is how to design the most efficient superuser route to achieve certain properties of end-to-end connectivity. Most existing solutions only focus on the network with stationary users or strongly constrained node mobility, and assume the superuser always moves with a fixed route. In this paper, we propose a flexible approach to design the superuser routes, considering the realistic user movements in MSNets. To the best of our knowledge, this work is the first to study active data dissemination from the social network perspective. We explore the geographic regularity of human mobility in the network, employ a semi-Markov analytical model to describe such mobility pattern, and hence formulate the superuser route design as a combinational optimization problem of Convex Optimization and Traveling Salesman Problem by exploiting social network concepts including communities and centrality. Extensive trace-driven simulations show that our approach consistently outperforms other existing superuser route design algorithms in terms of delivery ratio and energy efficiency.

Recent popularization of personal hand-held mobile devices makes it important to characterize the mobility pattern of mobile device users, so as to accurately predict user mobility in the future. Currently, the user mobility pattern is mostly characterized at a coarse-grained level, in the form of transition among wireless Access Points (APs). There is limited research effort on the fine-grained characterization of geographical user movement. In this paper, we present a novel approach to characterize the steady-state and transient-state user mobility behaviors at a fine-grained level, based on the Hidden Markov Model (HMM) formulation of user mobility. By applying our approach on both realistic mobility traces and synthetic mobility scenarios, we show that our approach is effective in characterizing user mobility pattern and making accurate mobility prediction. We also experimentally demonstrate that fine-grained user mobility knowledge is more effective to improve the performance of a variety of mobile computing applications.

The IP multimedia subsystem (IMS) is being deployed in the third generation (3G) networks since it supports many kinds of multimedia services. However, the security of IMS networks has not been fully examined. This paper presents a novel DoS attack against IMS. By congesting the presence service, a core service of IMS, a malicious attack can cause chained automatic reaction of the system, thus blocking all the services of IMS. Because of the low-volume nature of this attack, an attacker only needs to control several clients to paralyze an IMS network supporting one million users. To address this DoS attack, we propose an online early defense mechanism, which aims to first detect the attack, then identify the malicious clients, and finally block them. We formulate this problem as a change-point detection problem, and solve it based on the non-parametric GRSh test. Through trace-driven experiments, we demonstrate that our defense mechanism can throttle this DoS attack within a short defense time window while generating few false alarms.

Node mobility and end-to-end disconnections in Delay Tolerant Networks (DTNs) greatly impair the effectiveness of data dissemination. Although social-based approaches can be used to address the problem, most existing solutions only focus on forwarding data to a single destination. In this paper, we are the first to study multicast in DTNs from the social network perspective. We study multicast in DTNs with single and multiple data items, investigate the essential difference between multicast and unicast in DTNs, and formulate relay selections for multicast as a unified knapsack problem by exploiting node centrality and social community structures. Extensive trace-driven simulations show that our approach has similar delivery ratio and delay to the Epidemic routing, but can significantly reduce the data forwarding cost measured by the number of relays used.