Live video analytics requires multi-stage processing pipeline with low-latency that traditional cloud-centric approaches cannot always provide. While edge computing brings resources closer to users, edge nodes are resource-limited, heterogeneous, and carry dynamic workloads. Current load-balancing strategies often overlook key contextual factors such as communication overhead and compute pressure. Without full "foreseeable" context, systems suffer from latency spikes and high frame loss under bursty workloads or mobile deployments.

Indoor maps are essential for navigation, resource allocation, and autonomous operation in complex environments, yet creating them at scale has long been impeded by high costs and specialized hardware requirements. We present Tesseract, a modular system that transforms ordinary low-semantic floor plan images into navigable graph structures, without requiring specialized sensors or 3D modeling tools. Through Tesseract, we integrate deep learning modules for text detection and door classification. We then implement a novel floodfill-based segmentation and graph optimization solution. Tesseract ultimately generates semantically rich, compact graph representations of the original floor plans that are computationally parsable for indoor navigation applications. We evaluate Tesseract across two large-scale university buildings as well as a benchmark dataset, demonstrating high navigational completeness despite variations in layout complexity. The system processes floor plans efficiently, with runtime scaling linearly to the number of detected regions, thus remaining practical for large-scale deployments. Graph pruning reduces the initially dense connectivity—typically quadratic in the number of regions—to a sparse structure, yielding up to 78% fewer nodes and 70% fewer edges, all without compromising connectivity. Moreover, geometric fidelity is preserved within 80–86% of true real-world distances. These findings establish Tesseract as a robust and scalable solution, broadening access to automated indoor navigation and spatial analytics.

Brain-Computer Interfaces (BCIs), e.g., Neuralink, are evolving to enable seamless interaction and control across various domains, such as assistive systems and virtual environments. While conventional BCI applications treat users as control units that transmit commands to devices, our vision introduces a paradigm shift toward the novel concept of Human-as-a-Sensor (HaaS), where users themselves function as intelligent multimodal sensing agents. HaaS leverages BCIs to extract contextual spatial information from brain signals as individuals naturally interact with their environment. By decoding neural activity, HaaS complements the limitations of traditional sensors and offers unique advantages for spatially-aware systems. We explore a range of opportunities enabled by HaaS for enhancing spatial awareness and outline multi-disciplinary research challenges in realizing our vision. We also present a feasibility study on using HaaS to track human mobility, showcasing the promise of HaaS for powering future spatially-aware systems.

Recent transportation mode detection systems propose leveraging signals from only the serving cell tower to ensure ubiquity and practical deployability across all phones. However, existing solutions employ limited statistical hand-engineered features and traditional machine learning classifiers, leading to low estimation accuracy.

Designing a globally ubiquitous positioning system that works seamlessly in all environments remains a critical area of ongoing research. While Global Navigation Satellite Systems (GNSS), such as GPS, are the predominant technology for global outdoor positioning, they fail in areas with physical obstructions (e.g., dense urban regions and indoors) and are prone to jamming. These limitations significantly restrict their global accessibility and dependability, highlighting the need for supplementary positioning technologies.

All available cellular-based floor identification systems require information from multiple cell towers simultaneously, a feature absent in almost all phones, thus constraining their practicality. To address this limitation, we propose CellFloor, the first floor identification system to achieve high accuracy using only the serving cell tower while being regulatory-compliant. Based on recent advances in NLP, CellFloor builds a domain-specific edge-deployed large language model to identify the floor given a sequence of serving tower signal measurements. Our novel NLP-inspired approach allows CellFloor to extract rich contextual patterns from the signals, overcoming the limited information available when only the serving tower is used. Moreover, CellFloor employs recent advances in deep generative models to improve robustness against serving tower signal variations, enhancing floor identification accuracy. Our extensive evaluation of CellFloor shows consistently re-markable accuracy on multiple real testbeds, where it accurately estimates the exact floor at least 99.49% of the time using only the serving tower. This accuracy is superior to state-of-the-art (SOTA) systems, even when they use all available towers. Furthermore, unlike CellFloor, most SOTA systems fail to meet regulatory requirements when restricted to using just the serving tower, which is the only information available from the majority of current phones in the market. CellFloor also maintains regulatory compliance under different challenging conditions, including using only 20% of the available training data.

In spatial navigation, the shift from manual cartography to digital map representations has revolutionized how we interact with and comprehend outdoor and indoor environments. While digital mapping has substantially advanced outdoor navigation with robust techniques like satellite imagery and sophisticated data labeling, the full potential of indoor digital mapping remains untapped. Accurate indoor mapping promises to enhance the operational efficiency of mobile robots, improving their ability to interact with human environments, and bolstering emergency response capabilities. However, its realization is impeded by the complexity of current methods and the need for heavy manual labor, expert knowledge, and specialized equipment. To address these challenges, we introduce Text2Map – a novel methodology that harnesses natural language navigational instructions, the power of off-the-shelf Large Language Models (LLMs), and Few-shot Learning, to create graph-based digital maps of indoor spaces. This approach simplifies the mapping process for widespread use, leveraging crowd-sourceable ubiquitous navigation instructions as a data source without requiring specialized map data formats or hardware. Our paper presents the Text2Map system architecture, details the creation of the first dedicated dataset, and evaluates the system’s efficacy, highlighting the substantial potential and scalability of our approach. Text2Map achieves a Graph-Edit-Distance (GED) ranging from 0.5X to 2X the total number of regions in a building and an Edge Similarity score between 0.87 and 0.9. These results highlight the precision, robustness, and effectiveness of our methodology. Our work paves the way for a more accessible and streamlined approach to indoor digital mapping, setting the stage for broader adoption in human and mobile robot navigation applications.

Federated learning (FL) enabled creating models that are competitive to centralized machine learning models, without compromising user privacy. Participating FL clients train local models on their data and only share model weights. An FL server aggregates these weights into global weights that are pushed to clients for the next training round. Despite FL research growth, most of this work is conceived in experimental simulated environments that do not reflect its applicability to real-world scenarios. Also, existing open-source FL testbeds/frameworks have drawbacks that prohibit convenient deployment over a large spectrum of heterogeneous clients in realistic environments. These drawbacks include simulations, unrealistic data sets, not supporting heterogeneity, and not having realistic environment control in terms of network and client churn, for example. In this article, we introduce (RealFL) a novel, realistic, open-source, and extendable platform for FL that supports a large scale of heterogeneous clients. It enables a realistic assessment of FL solutions by controlling various environmental parameters, e.g., network, client churn, data distribution, training complexity, and client heterogeneity. Using these parameters, we assess RealFL performance through an extensive evaluation. Preliminary evaluation shows a performance gap of up to 72% in training time and 27% in accuracy between FL-simulated environments and RealFL. Moreover, extensive evaluation reveals that realistic environmental parameters could affect accuracy by up to 52.7%, training time by up to 77.5%, and communication overhead by up to 98%.

The continuous growth of Internet traffic, especially video content, presents challenges for access providers (APs) who must upgrade their infrastructure to meet increasing demands. Ensuring a high-quality experience (QoE) for end-users and finding ways to monetize network resources are key concerns. Guaranteeing QoE is complex, as it depends not only on link capacity but also on competing traffic flows and shared network data plane buffers. To address these challenges, we proposeDiffPerf, an in-network, online, and dynamic allocation system.DiffPerfoperates at both macroscopic and microscopic levels. At the macroscopic level, it elastically allocates bandwidth to performance-centric service classes defined by APs to accommodate different performance requirements. At the microscopic level,DiffPerfemploys a lightweight data-driven algorithm to statistically differentiate and isolate traffic flows within each class, improving their performance. We implementedDiffPerfprototypes using SDN-based technology, one with OpenDaylight and OpenFlow hardware switches, and the other with programmable Intel Tofino switches. Our evaluation focused on on-demand video streaming. The results demonstrate thatDiffPerfoffers APs a range of allocation choices while ensuring strong performance isolation. Additionally,DiffPerfimproves fairness and enhances overall user-perceived QoE within each class. Notably,DiffPerfconserves bandwidth and delivers a QoE improvement approximately$4.6\times $higher than TCP BBR, the most popular congestion control mechanism on the Internet.

Federated Learning is being hailed as a privacy-preserving machine learning alternative, by allowing models to be distributively trained on source devices owning their data. Most FL solutions, and their assessments, however, assume superior environmental reliability, despite the more realistic variances in environmental factors such as device and network capacity, data distribution, and device churn. As such, we argue in this paper, that there is a growing chasm between current FL assessment setups and the evolving FL assessment needs. Motivated by this chasm, we conduct, to the best of our knowledge, the first empirical measurement study of FL performance given realistic environmental factors. Our study quantifies the impact of these environmental factors on FL performance in terms of training time, accuracy, and communication overhead. Our findings have broad implications for the future development of FL including client admission control and scheduling optimizations.

Floor identification has gained much attention due to the increasing demand for indoor location-based services, especially prompt emergency response services. Leveraging Cellular signals for floor identification has been of recent interest due to the pervasiveness of cellular technology. However, all current systems rely on information from multiple cell towers concurrently, which is inaccessible in most phones and thus severely limits their deployability.

Federated Learning (FL) enabled creating models that are competitive to centralized Machine Learning models while preserving privacy by allowing clients to train data locally. Despite FL research growth, most of the work assessment and existing open-source FL testbeds/frameworks have drawbacks that prohibit convenient deployment over a large spectrum of heterogeneous clients in realistic environments. These drawbacks include simulations, unrealistic datasets, not supporting heterogeneity, and not having a realistic environment control in terms of network and client churn, for example. In this paper, we introduce (RealFL) a novel, realistic, open-source, and extendable platform for FL that supports a large scale of heterogeneous clients. It enables realistic assessment of FL solutions by controlling various environmental parameters; e.g. network, client churn, data distribution, training complexity, and client heterogeneity. Using these parameters, we assess RealFL performance through an extensive evaluation. The results show a performance gap of up to 77.5% in training time and 23.9% in accuracy between FL unrealistic environments and RealFL.

The need for a ubiquitous and accurate transportation mode estimation system has recently risen. Unfortunately, GPS-based and inertial sensor-based solutions lack this needed ubiquity and large-scale deployability, especially in developing countries. Thus, novel systems have proposed leveraging the more ubiquitous cellular technology. However, these systems either require cell tower locations or rely on information from multiple towers, which limits their deployability.We propose AutoSense, a ubiquitous and easily deployable transportation mode estimation system that works on all phones by relying on handover and received signal strength (RSS) information from only the serving cell tower. AutoSense offers a novel domain-specific deep learning-based system to perform automatic feature extraction and time-series processing. Our system handles several challenges, including limitations in cellular data, lack of location information, overfitting, and information decay in long-term dependencies. We extensively evaluate AutoSense using a real-world public dataset composed of 395 hours of data collected over seven months. Our results show that, compared to state-of-the-art systems, AutoSense can achieve enhancements in average precision and recall of 12.36% and 14.93%, respectively, while providing a highly ubiquitous and deployable solution using only the serving cell tower information.

Federated Learning (FL) has enabled Machine Learning (ML) applications to capture a larger spectrum of data by allowing such data to remain on-device, a desirable privacy guarantee in many applications. However, the highly iterative nature of FL optimization algorithms requires low-latency and high-throughput connections to clients. Unfortunately, realistic FL training scenarios include heterogeneous clients that are restricted by computation and communication, thereby slowing down or even failing FL training. In this paper, we propose FedTeams; a trust-based and resource-aware FL system that minimizes training latency, while improving accuracy. To achieve this, we mitigate the risk of straggling and weakly-connected clients by leveraging social trust and allowing these clients to offload their data to more powerful trusted peers that can train on their behalf. In specific, we formulate and solve an optimization problem that leverages the FedTeam’s trust graph and client resource information to optimize the distribution of training and minimize training latency. We evaluate FedTeams in a simulated environment, demonstrating up to a 81.6% decrease in training latency and 11.2% increase in global model accuracy when compared to existing state-of-the-art solutions.

Earables have been increasingly gaining attention from consumers and manufacturers alike due to their small footprint, ease of use, and the added accessibility they bring. However, the limited interface of these devices, usually being a single button or force-sensor, inhibits their potential. Earables can provide a much richer experience by extending the ways in which users can interact with them. In this paper, we present EarGest, a novel earable-based hand gesture recognition system that does not require calibration or training, and works with commercially available BLE-enabled earphones and devices. Our proposed system is unique in its ability to leverage Bluetooth to detect hand motion near the ear to recognize gestures. By harnessing information from BLE connections between the wireless earphones and a host device, we accurately detect and classify different hand gestures performed by users, while also determining discrete levels of hand speed. EarGest operates without interfering with the regular functionality of the earphones and introduces minimal energy overhead on the host device. We implement a prototype of the system using eSense, a multi-sensory earable platform, and evaluate it in different scenarios and settings. Results show that our system can detect and classify seven near-ear hand gestures with an accuracy up to 98.5%, as well as identify hand motion speed with 96% accuracy.

Latency and privacy concerns, together with the spread of IoT devices, have recently sparked interest in edge computing and computational offloading to the edge and beyond. Portability and migratability are important requirements to achieve a stable edge femtocloud offloading platform. Due to the inherent heterogeneity of the edge, code compatibility is one of the core challenges towards achieving those goals. In this paper, we examine how popular technologies achieve portability and migratability, and the advantages and limitations of each method in light of established standards and current research. We argue that WebAssembly has many advantages as a platform for femtocloud offloading. Next, we implement Nomad, an interpreter-based environment, to run WebAssembly that is capable of live-migrating across operating systems and hardware architectures. We evaluate Nomad and examine the effect of migration on performance, startup and migration delays, and cross-platform migration. We find that migration, even cross-platform, adds less than 5% overhead to performance and fixed delays are less than 2ms in most scenarios.

Mobile and IoT devices are becoming increasingly capable computing platforms that are often underutilized. In this paper, we propose RAMOS, a system that leverages the idle compute cycles in a group of heterogeneous mobile and IoT devices that can be clustered to form an edge FemtoCloud. At the heart of this system, we formulate a multi-objective, resource-aware task assignment and scheduling problem. The scheduler runs in two main modes; latency-minimization and energy-efficiency. Under the latency-minimization mode, it strives to maximize the computational throughput of the constructed FemtoCloud while maintaining the energy consumption below an operator specified threshold. Under the energy-efficient mode, it minimizes the total energy consumed in the FemtoCloud while meeting defined tasks deadlines. Due to the NP-Completeness of this scheduling problem, we design a set of heuristics to solve it. We implement a prototype of our system and use it to evaluate its performance and efficiency. Our results demonstrate the system's ability to meet different scheduling objectives while adhering to pre-specified time and energy constraints. Compared to other schedulers, RAMOS achieves 10 to 40 percent completion time improvement under latency minimization mode and up to 30 percent more energy-efficiency under the energy-efficient mode.

Many novel IoT-based applications now require large compute resources, high-privacy, and low-latency. This demand has triggered the rise of fog and edge computing to complement the high-latency and low-privacy cloud. Fog computing provides lower latency by bringing computational servers closer to the user, typically within the city's vicinity. However, due to the high cost of deploying such fog servers at scale, and poor network infrastructures in many countries and areas, edge computing has been introduced. Edge computing argues for leveraging compute resources, typically within a user's immediate environment, on distributed ensembles of devices called FemtoClouds. In this paper, we propose Maestro, a system that aids users by offloading computational jobs from them to multiple FemtoClouds in their immediate vicinity. We propose an integrated architecture for Maestro, which incorporates a new scheduling algorithm that assigns compute tasks to FemtoClouds. We implement a full prototype of Maestro, and evaluate its performance on our experimental testbed, as well as through emulation. Our results show that our system and scheduler outperforms state-of-the-art by up to 55%.

The extreme growth and diversity of IoT applications, along with the heterogeneity of these devices, has led to numerous middleware solutions emerging to address various relevant challenges. These solutions have been shifting from cloud-based approaches to edge-technologies in order to handle issues related to privacy, latency, and bandwidth. Within this context, we identify an opportunity to introduce a novel IoT Middleware system that enables more seamless resource and context sharing at the edge. We propose the Hive, a middleware system that allows heterogeneous off-the-shelf devices in a common edge network to seamlessly and efficiently utilize each other's sensory and computational resources. The Hive architecture enables a new wave of multi-modal sensory applications, leveraging a pool of IoT devices, that would otherwise be unattainable. We accomplish this by decoupling the hardware, processing, and application layers within IoT devices from each other. The Hive abstracts each of these layers into a single resource pool that is shared and cross utilized on-demand within the edge network by any individual device. We implement the Hive, along with a dedicated communication protocol for our system. We evaluate the effectiveness of the Hive by integrating it with two sample IoT applications: an audio-based emotion recognition system, and a video-based facial detection application. We extensively evaluate the impact the Hive has on these new applications after integration, and additionally investigate its impact at scale. Our results show how the hive boosts the overall utilization of resources in an edge IoT network, reduces computational delay in complex applications, and most importantly, enables applications to perform at higher level of effectiveness.

Having knowledge of the users' indoor location and the semantics of their environment can facilitate the development of many indoor context-aware applications. For such applications, an accurate indoor map is often needed. While current techniques are capable of producing such maps, these maps are not labeled and hence are of limited utility for many applications. To address this shortcoming, we propose Sherlock, a crowdsourced system for automatically tagging indoor floor plans. Sherlock leverages the myriad of sensors embedded in modern smartphones to intelligently gather audio and visual data, and upload it to the Sherlock Server. At the Sherlock Server, acoustic monitoring and object recognition techniques are used to classify these data samples. The classification scores of current and past samples are then aggregated in a probabilistic framework to determine the confidence with which we can apply as label to a given space. We evaluate Sherlock on a dataset of more than 11,000 audio recordings and 1,200 images, that we collected in three different university campuses. In our evaluation, the confidence for the true label generally outstripped the confidence for all other labels and, in some cases, even reached as high as 100% with as little as 30 data samples.

WiFi time of flight (ToF) measurement has been supported recently by the wireless LAN protocols to improve WiFi localization. Specifically, the IEEE 802.11-2016 standard has a fine-time measurement (FTM) protocol that can be used to measure the WiFi signal round trip time (RTT). In this paper, we present the design and implementation of WiNar, a WiFi RTT-based indoor location determination system that combines the advantages of both fingerprint and ranging-based techniques to overcome the different challenges of indoor environments. Using commercial-off-the-shelf access points and mobile phones, WiNar leverages the propagation time of the wireless signal with a fingerprinting model to address the multipath, non-line-of-sight, signal attenuation, and interference challenges of the indoor environments. Moreover, when leveraging the round trip time measurements, WiNar does not require clock synchronization between the transmitter and the receiver. We discuss the different components of the system and its implementation on the Android operating system. Our results show that WiNar has a sub-meter localization accuracy with an average localization error of less than 0.86 meters for two different testbeds. This accuracy outperforms the performance of the traditional signal strength (RSS) fingerprinting technique by at least 38% and ranging-based multi-lateration technique by at least 148%. Finally, our system is also robust to heterogeneous devices.

We present MagStroke, a system that provides a more flexible mechanism for providing textual input to smart devices with small screens. The key idea is that MagStroke leverages the effect of a magnet worn on the user's hand on a nearby device's magnetometer to enable the use of larger virtualized keyboards. While typing a certain key using a virtual standard keyboard, the hand and fingers of the user, with the attached magnet, move in a unique motion and thus a unique pattern is formed in the device's embedded magnetometer sensor readings. To realize MagStroke, we address several challenges including handling background magnetic noise, keystroke detection and pattern segmentation, feature selection, and training overhead reduction. We implement a proof-of-concept prototype using off-the-shelf mobile devices and extensively evaluate the different modules in our system. Our results show that MagStroke can achieve a word recognition accuracy of 95% with a 46% reduction in training overhead, highlighting its promise as a ubiquitous off-the-shelf magnetic-based keyboard.

We propose DeepNar, a deep learning-based indoor localization system that leverages WiFi signals time of flight (ToF) as environment features, to provide accurate and robust indoor localization. DeepNar leverages the fine-time measure¬ment (FTM) protocol in the recent IEEE 802.11-2016 standard to measure WiFi signal round trip time (RTT). Our system combines the advantages of fingerprinting and ranging-based techniques by providing a deep learning model along with a probabilistic framework that captures the complex relation between the propagation times of the WiFi signals heard by the mobile phone and its location. By leveraging the signals RTT, collected using commercial-off-the-shelf access points and mobile phones, DeepNar overcomes the different challenges of indoor environments such as the multipath interference, non-line-of-sight transmissions, signal attenuation, and interference. Moreover, DeepNar does not require clock synchronization between the transmitter and the receiver. Our system is composed of various components that handle outlier detection, avoids over-training, and accommodates heterogeneous devices. We implement and evaluate DeepNar over two testbeds. Our results show that DeepNar has a sub-meter localization accuracy with a median error less than 0.75m. This accuracy outperforms ranging-based multi-lateration technique by at least 182% and traditional signal strength (RSS) fingerprinting techniques by more than 119% and 33% in both testbeds considered.

Fog and cloud computing are becoming increasingly popular as computational offloading platforms. However, many areas throughout the world either do not have a stable enough network connection to effectively use cloud computing, or do not have the required infrastructure or resources to set up fog nodes. Moreover, many application categories such as cognitive assistance and augmented reality applications require single digit latency, which both cloud and fog cannot provide. As such, these areas can benefit tremendously from edge computing. To ensure single digit latency, Edge nodes are typically other co-located devices within single-hop proximity. Having said that, due to user mobility, single edge devices set up at home will not provide the required latency if the user is on the move. Therefore, we introduce Teddybear, a Docker based system that seamlessly and efficiently migrates server containers between edge computing platforms by utilizing both the Internet and the user's mobile device as a carrier for the container. We show how Teddybear can continue to provide ultra-low latency services to users on the move in areas with poor network bandwidths/infrastructures, with minimal service downtime as they change locations.

Although high speed internet access is widespread in many countries, it is still undeniably scarce elsewhere. Opportunistic networking solutions have provided alternative connectivity models in such scenarios. One of the main opportunistic networking approaches involves leveraging social and Space Syntax information in order to minimize message forwarding costs, along with delay. Space Syntax is a domain of architecture engineering that relates spatial configuration to user mobility and thus, can provide insights into when to forward messages. In this work, we question the accuracy of existing Space syntax metrics in defining attraction points in a given urban area, and statistically demonstrate their inefficiency and lack of power awareness. We propose variations of Space Syntax metrics, and introduce a framework for integrating Space Syntax with social-based forwarding algorithms. We evaluate our framework via simulations using real traces that we collected from a university campus. Our results show an improvement in performance resulting from this integration, in terms of f-measure and utilization fairness, as well as significant reduction in the ratio of contacted uninterested nodes, cost, power consumption, and delay.

Promoting smart and scalable remote health monitoring systems is challenging due to the enormous amount of collected data that needs to be processed and transferred given the limited network resources and battery-operated devices. Thus, the conventional cloud computing paradigm alone, is not always the most suitable solution for enabling such systems. In this context, we propose and implement a smart edge-based health system that aims at decreasing the system latency and energy consumption, while optimizing the delivery of the medical data. In particular, we formulate a multi-objective optimization framework that enables an edge node to dynamically adjust compression parameters and select the optimal radio access technology (RAT) while maintaining a trade-off between energy consumption, latency, and distortion. Furthermore, to evaluate and verify our framework, we develop an experimental testbed, where a data emulator is implemented to send EEG data to an edge node that classifies, compresses, and transfers the gathered data through the optimal RAT to the health cloud. Our experimental results show that the proposed system can offer about 30% energy savings while decreasing the delivery time to half of its value compared to a system that lacks edge processing capabilities.

Fifth generation mobile communication networks are currently being deployed, thus making Tactile Internet possible. Tactile Internet is the future advancement of the current Internet of Things (IoT) vision wherein haptics, or touch and senses, can be communicated from one geographical place to another, enabling near real-time control and navigation of remote objects. Tactile Internet will have its use cases in several application domains, with the industrial sector being among the most prominent ones. With the Industrial Internet of Things (IIoT), Tactile Internet will be used in healthcare, manufacturing, mining, education, autonomous driving, etc. The acceptable delay in most of these tactile applications will be under one millisecond. Since Tactile Internet communicates haptics and gives visual feedback, quality of service (QoS) becomes an important issue. Similarly, user's satisfaction on the service quality [often measured as quality of experience (QoE)] becomes equally important. To reap the true potential of Tactile Internet, sophisticated and intelligent mechanisms are required between the end-nodes. A middleware such as fog computing can be vital in this context, since it can allocate resources based on the QoS/QoE requirements of each service. In this context, we present a QoE-aware model for dynamic resource allocation for tactile applications in IIoT. We implement the model using Java and discuss the empirical results to elaborate more on the impact of such a model for QoE-aware resource allocation that can be very important in the context of Tactile Internet, especially IIoT. We also discuss some of the most prominent use cases of Tactile IIoT.

We present WiGest: a system that leverages changes in WiFi signal strength to sense in-air hand gestures around the user's mobile device. WiGest uses standard WiFi equipment, with no modifications, and requires no training for gesture recognition. The system identifies different RSS change primitives, from which we construct mutually-independent gesture families. These families can be mapped to distinguishable application actions. More fine-grained features can also be recognized for the detected primitives using CSI. WiGest addresses various challenges including cleaning the noisy signals, gesture type and attribute detection, reducing false positives due to interfering humans, and adapting to changing signal polarity. We implement a proof-of-concept prototype using off-the-shelf devices and extensively evaluate the system in two different environments. Our results show that WiGest detects the basic primitives with an accuracy of 87.5 percent using one AP, including through-the-wall non-line-of-sight scenarios, which increases to 96 percent using three overheard APs. Additionally, when evaluating the system using a multi-media player application, we achieve an accuracy of 96 percent. This accuracy is robust to the presence of other interfering humans, highlighting WiGest's ability to enable future ubiquitous hands-free gesture-based interaction with mobile devices.

Routing protocols in multi-hop cognitive radio networks (CRNs) can be classified into two main categories: local and global routing. Local routing protocols aim at decreasing the overhead of the routing process while exploring the route by choosing, in a greedy manner, one of the direct neighbors. On the contrary, global routing protocols choose the optimal route by exploring the whole network to the destination paying the flooding overhead cost. In this paper, we propose a primary user-aware$k$-hop routing scheme where$k$is the discovery radius. This scheme can be plugged into any CRN routing protocol to adapt, in real time, to network dynamics like the number and activity of primary users. The aim of this scheme is to cover the gap between local and global routing protocols for CRNs. It is based on balancing the routing overhead and the route optimality, in terms of primary users avoidance, according to a user-defined utility function. We analytically derive the optimal discovery radius ($k$) that achieves this target. Evaluations on NS2 with a side-by-side comparison with traditional CRNs protocols show that our scheme can achieve the user-defined balance between the route optimality, which in turn reflected on throughput and packet delivery ratio, and the routing overhead in real time.

Low-cost mini-drones with advanced sensing and maneuverability enable a new class of intelligent sensing systems. To achieve the full potential of such drones, it is necessary to develop new enhanced formulations of both common and emerging sensing scenarios. Namely, several fundamental challenges in visual sensing are yet to be solved including (1) fitting sizable targets in camera frames; (2) positioning cameras at effective viewpoints matching target poses; and (3) accounting for occlusion by elements in the environment, including other targets. In this article, we introduce Argus, an autonomous system that utilizes drones to collect target information incrementally through a two-tier architecture. To tackle the stated challenges, Argus employs a novel geometric model that captures both target shapes and coverage constraints. Recognizing drones as the scarcest resource, Argus aims to minimize the number of drones required to cover a set of targets. We prove this problem is NP-hard, and even hard to approximate, before deriving a best-possible approximation algorithm along with a competitive sampling heuristic which runs up to 100× faster according to large-scale simulations. To test Argus in action, we demonstrate and analyze its performance on a prototype implementation. Finally, we present a number of extensions to accommodate more application requirements and highlight some open problems.

The recent growth in smart phone adoption coupled with social networking has lead to an increase in user generated content (UGC). The majority of UGC, captured in the form of images, videos, and audio clips, have significantly grown in size as a result of advancements in multimedia technology with higher definition phone cameras. Online clouds are generally the go-to solution in order to accommodate this increase in storage requirements. However, online clouds raise privacy concerns, are not fully automated, and do not adapt to different networking environments. We propose an edge based automated storage management system, EdgeStore, that utilizes user-owned devices at the edge, to automatically offload and retrieve files. EdgeStore determines file popularity based on the user's access patterns and offloads unpopular files. It accounts for different networking infrastructures, ranging from rural areas to metropolitan areas, in order to serve users in different environments. We implement and evaluate EdgeStore showing that users can offload up to 90% of storage to underutilized edge devices, and still have a high percentage of low latency access to offloaded files within 10 seconds.

The basic MAC mechanisms in IEEE 802.11 (WiFi) have remained largely unchanged for over 20 years. In this paper, we argue that the prevalence of WiFi makes it almost impossible to improve its performance through changes that require modifying hardware, firmware, or drivers. New applications, however, continue to exert novel performance demands. We suggest that changes should be developed as augmentation-only solutions through above-driver, kernel-level software modifications. An augmentation-only solution needs to maintain inter-operability and afford transparency in performance to existing WiFi devices, as well as enable minimum overhead upgradability. Our goal is to demonstrate the feasibility of MAC augmentation according to these principles. To this end, we leverage soft scheduling, where nodes are asked for a best-effort attempt to adhere to a given schedule. We allow the soft scheduler to coexist with and work at a different time scale from WiFi's Distributed Coordination Function (DCF); allowing it to reduce the time nodes spend contending for the medium while allowing DCF to handle only missed schedule slots and schedule divergence. We present a new Soft Token Passing Protocol (STPP) as an instance of this family of Soft Scheduling Protocols. We then show how STPP can be made part of a MAC protocol with specific performance improvement goals by developing the Wireless Low-Latency Local Links (WL4) system. We evaluate WL4 on a five node microbenchmark and quantify the system's overhead on network throughput and latency. We show that soft scheduling, via STPP, enables WL4 to adhere to our augmentation principles while improving the latency within the system.

CAPTCHA challenges are used all over the Internet to prevent automated scripts from spamming web services. However, recent technological developments have rendered the conventional CAPTCHA insecure and inconvenient to use. In this paper, we propose vCAPTCHA, a voice-based CAPTCHA system that would: (1) enable more secure human authentication, (2) more conveniently integrate with modern devices accessing web services, and (3) help collect vast amounts of annotated speech data for different languages, accents, and dialects that are under-represented in the current speech corpora, thus making speech technologies accessible to more people around the world. vCAPTCHA requires users to speak their responses, in order to unlock or use different web services, instead of typing them. These user responses are analyzed to determine if they were indeed naturally produced, and transcribed to ensure that they contain the challenge sentence. We build a prototype for vCAPTCHA in order to assess its performance and practicality. Our preliminary results show that we are able to achieve an attack success rate as low as 2.3% while maintaining a human success rate comparable to current CAPTCHAs, on ASVspoof datasets.

In the context of mobile Health (mHealth) applications, data are prone to several sources of contamination which would lead to false interpretation and misleading classification results. In this paper, a robust deep learning approach with low rank model is proposed to classify mHealth vital signs. Further-more, we propose using the Schatten-p norm instead of the classic nuclear norm since it has shown better recovery performance for several applications. We conduct a comprehensive study where we compare our method to the state-of-art methods and evaluate its performance with respect to the key system parameters. Our findings show indeed that combining deep network with dictionary learning model is effective for vital signs classification even in presence of 50% corruption with 8% improvement over the closest performance.

Rapid technological advances have revolutionized the industrial sector. These advances range from automation of industrial processes to autonomous industrial processes, where a human input is not required. Internet of Things (IoT), which has emerged a few years ago, has been embraced by industry, resulting in what is known as the Industrial Internet of Things (IIoT). IIoT refers to making industrial processes and entities part of the Internet. Restricting the definition of IIoT to manufacturing yields another subset of IoT, known as Industry 4.0. IIoT and Industry 4.0, will consist of sensor networks, actuators, robots, machines, appliances, business processes, and personnel. Hence, a lot of data of diverse nature would be generated. The industrial process requires most of the tasks to be performed locally because of delay and security requirements and structured data to be communicated over the Internet to web services and the cloud. To achieve this task, middleware support is required between the industrial environment and the cloud/web services. In this context, fog is a potential middleware that can be very useful for different industrial scenarios. Fog can provide local processing support with acceptable latency to actuators and robots in a manufacturing industry. Additionally, as industrial big data are often unstructured, it can be trimmed and refined by the fog locally, before sending it to the cloud. We present an architectural overview of IIoT and Industry 4.0. We discuss how fog can provide local computing support in the IIoT environment and the core elements and building blocks of IIoT. We also present a few interesting prospective use cases of IIoT. Finally, we discuss some emerging research challenges related to IIoT.

Mobile devices are getting increasingly popular all over the world. Mobile users in developing countries however rarely have Internet access which puts them at economic and social disadvantages compared to their counterparts in developed countries. We propose mBridge: A distributed system to disseminate multimedia content to mobile users with intermittent Internet access and opportunistic ad hoc connectivity. By disseminating various multimedia content such as news reports notification messages targeted advertisements movie trailers and TV shows mBridge aims to eliminate the digital divide. We formulate an optimization problem to compute personalized distribution plans for individual mobile users to maximize the overall user experience under various resource constraints. Our formulation jointly considers the characteristics of multimedia content mobile users and intermittent networks. We present an efficient distribution planning algorithm to solve our problem and we develop several online heuristics to adapt to the system and network dynamics. We implement a prototype system and demonstrate that our algorithm outperforms the existing algorithms by up to 206% 472% and 188% in terms of user experience disk efficiency and energy efficiency respectively. In addition we conduct trace-driven simulations to rigorously evaluate the proposed system in different environments and for large-scale deployments. Our simulation results demonstrate that the proposed algorithm substantially outperforms the closest ones in the literature in all performance measures. We believe that mBridge can allow multimedia content providers to reach out to more mobile users and mobile users to access multimedia content without always-on Internet access.

With the advancements in mobile sensors and health-care technologies, mobile health (mHealth) services are growing in demand. However, the deployment of mHealth's applications in rural and underdeveloped areas remains a major challenge, despite the investments made, largely due to unreliable communication infrastructures. In this paper, we propose delivering mHealth data in a highly disruptive wireless network using resource-limited mobile devices. We build on state-of-the-art opportunistic/DTN solutions, and propose two dynamic schemes that adapt to the level of congestion in the network. Our adaptive forwarding schemes dynamically tune data replication at forwarder nodes by engaging the most appropriate forwarding strategy at any given state, while incurring minimal overhead. In order to achieve such a goal, we propose a reactive and proactive approach to detecting or predicting congestion in the network respectively. We perform a set of data-driven simulations to compare the performance of our proposed schemes with state-of-the-art DTN forwarding algorithms. Our results show that our schemes achieve better delivery ratio for realistic mHealth applications in challenged networking environments.

We introduce a mobile sensing technique to detect a nearby active television, the channel it is tuned to, and whether it is receiving this channel over the air or not. This technique can find applications in tracking TV viewership, second screen services and advertising, as well as improving the efficiency of TV white space spectrum usage. The technique uses a three-stage detection process: It first uses a Gaussian mixture model on audio recordings from mobile phones to detect likely TV sounds in the area. It then correlates the recording with known TV channel audio to identify the channel and improve detection robustness. Finally, it applies a latency analysis to determine whether programming is received over-the-air or through alternate means such as cable or satellite TV. Our system is evaluated using diverse datasets that take into account different realistic scenarios of indoor environments for several users. The results show that the system can achieve an area under the curve (AUC) of 0.9979 and a false negative rate of 0.0132.

White spaces are portions of the TV spectrum that are allocated but not used locally. Ifaccurately detected, white spaces offer a valuable new opportunity for highspeed wireless communications. We propose a new method for white space detection that allows a node to actlocally, based on a centrally constructed model, and at low cost, whiledetecting more spectrum opportunities than best known approaches. Weleverage two ideas. First, we demonstrate that low-cost spectrum monitoringhardware can offer "good enough" detection capabilities. Second, we develop amodel that combines locally-measured signal features and location to more efficiently detect white space availability. We incorporate these ideas into the design,implementation, and evaluation of a complete system we call Waldo. We deployWaldo on a laptop in the Atlanta metropolitan area in the US covering 700 km2. Our results show that usingsignal features, in addition to location, can improve detection accuracy by up to10x for some channels. We also deploy Waldo on an Android smartphone,demonstrating the feasibility of real-time white space detection with efficientuse of smartphone resources.

Low-cost mini-drones with advanced sensing and maneuverability enable a new class of intelligent visual sensing systems. This potential motivated several research efforts to employ drones as standalone surveillance systems or to assist legacy deployments. However, several fundamental challenges remain unsolved including: 1) Adequate coverage of sizable targets; 2) Target orientation that render coverage effective only from certain directions; 3) Occlusion by elements in the environment, including other targets.

The proliferation of mobile health (mHealth) technologies in rural and disconnected areas has enabled novel communication challenges. These challenges exist due to the underdeveloped Internet infrastructure, wireless intermittent connectivity, and the large data generated by such applications that often overwhelms the weak infrastructure. In this paper, we investigate data delivery in such overloaded and sparse rural area scenarios, where patients deploy an mHealth service over a disruption-tolerant network (DTN). To quantify the feasibility and the challenges of such communication scenario, we investigate the performance of multiple opportunistic forwarding algorithms under different environments and loads. We highlight the shortcoming of most of these algorithms with regards to the lack of efficient resource management algorithms and the large overhead introduced by such state-of-the-art algorithms. Our data driven experiments highlight such limitations while comparing the performance of three classes of algorithms; flooding based, resource-aware, and controlled flooding algorithms.

We present an empirical study for device-to-device single hop wireless communication. We perform realistic indoor and outdoor measurements of WiFi and Bluetooth communication between representative mobile and IoT devices. This measurement paper aims at answering multiple questions: What are the methods and the limitations in D2D wireless communications for multiple IoT devices? What is the impact of distance/obstacles, environment, and device types on TCP and UDP data transfers? We address these questions, and more, by conducting experimental measurements to cover varying performance aspects such as effective throughput, packet loss ratio, received signal strength (RSSI), and end-to-end delay. Our study presents insights and awareness of the true communication performance of different IoT devices and wireless communication technologies.

In this paper, we present a joint compression and classification approach of EEG and EMG signals using a deep learning approach. Specifically, we build our system based on the deep autoencoder architecture which is designed not only to extract discriminant features in the multimodal data representation but also to reconstruct the data from the latent representation using encoder-decoder layers. Since autoencoder can be seen as a compression approach, we extend it to handle multimodal data at the encoder layer, reconstructed and retrieved at the decoder layer. We show through experimental results, that exploiting both multimodal data intercorellation and intracorellation 1) Significantly reduces signal distortion particularly for high compression levels 2) Achieves better accuracy in classifying EEG and EMG signals recorded and labeled according to the sentiments of the volunteer.

Due to the exponential advancements in mobile and IoT device capabilities, we are currently witnessing a resurrection in the domains of ubiquitous and pervasive systems. These systems and devices heavily rely on efficient wireless communication as an enabler of single and multi-hop wireless data transfers. Awareness of the practical performance capabilities and limitations of existing wireless technologies in various heterogeneous settings is fundamental for the deployment of effective ubiquitous applications. We conduct and analyze an in-depth and detailed single-hop and various multi-hop networking measurements study utilizing a representative set of emerging mobile and IoT devices equipped with a variety of wireless interfaces. Additionally, given the popularity of Android-based D2D systems developed, we propose two solutions that would overcome the inherent two-hop D2D WiFi-Direct limitation, and provide many-hop WiFi-Direct-based communication. Our results provide deep insights on the performance of multi-hop D2D communication within various environments, along with the limitations and trade-offs in the communication technologies, topologies, and solutions adopted.

Following the explosion of Internet-enabled mobile devices and mobile services, users tend to expect network connectivity everywhere they go and at any time. In anticipation of these higher connectivity expectations and requirements especially for vehicular communication, several technologies like WiMAX, cellular networks (i.e., 3G/4G) and wireless mesh networks have been developed. While cellular network communication is either costly or slow, WiMAX which provides data rates comparable to regular WiFi networks, suffers from capacity issues as there are limited channels and hence a limited number of users able to transmit to a single base station. This paper investigates the connectivity challenges in WiFi vehicular opportunistic communications and discusses the utility of such networks at various speeds. We perform a set of simulations to compare state-of-the-art solutions that aim at enhancing the access point entry, production, and exit phases of such communication. We also emulate a best-case solution that enhances most state-of-the-art communication features in order to quantify the potential improvement that can be achieved in vehicular opportunistic networks. The ultimate goal is to provide insights into the true promise and potential of vehicular opportunistic networking.

One of the main methods for interacting with mobile devices today is the error-prone and inflexible touch-screen keyboard. This paper proposes MagBoard: a homomorphic ubiquitous keyboard for mobile devices. MagBoard allows application developers and users to design and print different custom keyboards for the same applications to fit different user's needs. The core idea is to leverage the triaxial magnetometer embedded in standard mobile phones to accurately localize the location of a magnet on a virtual grid superimposed on the printed keyboard. This is achieved through a once in a lifetime fingerprint. MagBoard also provides a number of modules that allow it to cope with background magnetic noise, heterogeneous devices, different magnet shapes, sizes, and strengths, as well as changes in magnet polarity. Our implementation of MagBoard on Android phones with extensive evaluation in different scenarios demonstrates that it can achieve a key detection accuracy of more than 91% for keys as small as 2cm × 2cm, reaching 100% for 4cm×4cm keys. This accuracy is robust with different phones and magnets, highlighting MagBoard promise as a homomorphic ubiquitous keyboard for mobile devices.

Unmanned Aerial Vehicles are miniature air-crafts that have proliferated in many military and civil applications. Their affordability allows for tasks to be held with not just one but a fleet of UAVs. One of the problems that arise with the use of multi-UAVs is the multi-UAV path planning and assignment problem. We propose three algorithms that aim at assigning energy efficient trajectories for a fleet of UAVs. Our optimal path planning solution (OPP) is formulated using a Mixed Integer Linear Programming model (MILP). We also propose two other heuristic solutions that are greedy in nature; namely, Greedy Least Cost (GLC) and First Detect First Reserve (FDFR). To aid with collision avoidance, we adopt the concept of space discretization, and present a more realistic view of the space a UAV occupies. The comparative study of our proposed solutions reveals insightful trade-offs between energy consumption and complexity.

Many applications are being developed to leverage the popularity of mobile opportunistic networks. However, building adaptive testbeds can be costly and challenging. This challenge motivates the need for effective opportunistic network simulators to provide a variety of opportunistic environment setups, and evaluate proposed applications and protocols with a comprehensive set of metrics. This paper presents SAROS, a simulator of opportunistic networking environments with a variety of interest distributions, power consumption distributions, imported real traces, and social network integration. The simulator provides a wide variety of evaluation metrics that are not offered by comparable simulators. Finally, SAROS also implements several opportunistic forwarding algorithms ranging from social-oblivious algorithms to interest and power-aware social-based algorithms.

Mobile devices are becoming increasingly capable computing platforms with significant processor power and memory. However, mobile compute capabilities are often underutilized. In this paper we consider how a collection of co-located devices can be orchestrated to provide a cloud service at the edge. Scenarios with co-located devices include, but are not limited to, passengers with mobile devices using public transit services, students in classrooms and groups of people sitting in a coffee shop. To this end, we propose the femtocloud system which provides a dynamic, self-configuring and multi-device mobile cloud out of a cluster of mobile devices. We present the femtocloud system architecture designed to enable multiple mobile devices to be configured into a coordinated cloud computing service despite churn in mobile device participation. We develop a prototype of our femtocloud system and use it in addition to simulations to evaluate the performance of the system showing its efficiency and ability to leverage the available devices' compute capacity. We contribute to a line of research on small, local and possibly private clouds.

With the advent of wearable computing and the resulting growth in mobile application market, we investigate mobile opportunistic cloud computing where mobile devices leverage nearby computational resources in order to save execution time and consumed energy. Our goal is to enable generic computation offloading to heterogeneous devices that include Cloud, mobile devices, and cloudlets. We propose a generic and flexible architecture that maximizes the computation gain with respect to various objective functions such as, minimizing the response time, reducing the overall energy consumption, and increasing the network lifetime. This novel architecture is designed to automate computation offloading to numerous compute resources over disrupted network connections.

The evolution of mobile devices into highly capable computing platforms that sense, store, and execute complex tasks is making them attractive candidates for edge computational micro-cloud settings. Such solutions are creating novel security challenges due to the increased push for more seamless computational cyber-foraging that leverages the exploding proliferation of mobile devices. A major concern is that security challenges stemming from these trends, are growing at a rate exceeding the evolution of security solutions. In this paper, we consider an environment in which computational offloading is performed among a set of mobile devices. We propose HoneyBot, a defense technique for device-to-device (d2d) malicious communication. While classical honeypots designed to isolate distributed denial of service (DDoS) botnet attacks fail to detect d2d insider attacks, HoneyBot nodes detect, track, and isolate such attacks. We propose and investigate detection and tracking algorithms that leverage insecure d2d infected communication channels to accurately and efficiently identify suspect malicious nodes and isolate them. Our data driven evaluation and analysis, based on 3 real world mobility traces, show that the number and placement of HoneyBot nodes (Hb) in the network considerably impact the tracking delay and the detection accuracy.

We propose a primary user-aware k-hop routing scheme that can be plugged into any cognitive radio network routing protocol to adapt, in real time, to the environmental changes. The main use of this scheme is to make the compromise required between the route overhead and its optimality based on a user-defined utility function. We analytically derive the optimal discovery radius (k) that achieves this target. Evaluations on NS2 show that our scheme can enhance the current routing protocols in terms of throughput with minimal overhead.

We present WiGest: a system that leverages changes in WiFi signal strength to sense in-air hand gestures around the user's mobile device. Compared to related work, WiGest is unique in using standard WiFi equipment, with no modifications, and no training for gesture recognition. The system identifies different signal change primitives, from which we construct mutually independent gesture families. These families can be mapped to distinguishable application actions. We address various challenges including cleaning the noisy signals, gesture type and attributes detection, reducing false positives due to interfering humans, and adapting to changing signal polarity. We implement a proof-of-concept prototype using off-the-shelf laptops and extensively evaluate the system in both an office environment and a typical apartment with standard WiFi access points. Our results show that WiGest detects the basic primitives with an accuracy of 87.5% using a single AP only, including through-the-wall non-line-of-sight scenarios. This accuracy increases to 96% using three overheard APs. In addition, when evaluating the system using a multi-media player application, we achieve a classification accuracy of 96%. This accuracy is robust to the presence of other interfering humans, highlighting WiGest's ability to enable future ubiquitous hands-free gesture-based interaction with mobile devices.

The exponential increase in mobile data demand, coupled with growing user expectation to be connected in all places at all times, have introduced novel challenges for researchers to address. Fortunately, the wide spread deployment of various network technologies and the increased adoption of multi-interface-enabled devices allow researchers to develop solutions for those challenges. Such solutions exploit available interfaces on these devices in both local and collaborative forms. These solutions, however, have faced a formidable deployment barrier. Therefore, in this paper, we present OSCAR, a multi-objective, incentive-based, collaborative, and deployable bandwidth aggregation system, designed to exploit multiple network interfaces on modern mobile devices. Oscar's architecture does not introduce any intermediate hardware nor require changes to current applications or legacy servers. This architecture estimates the interfaces characteristics and application requirements, schedules various connections and/or packets to different interfaces, and provides users with incentives for collaboration and bandwidth sharing. We formulate the OSCAR scheduler as a multi-objective scheduler that maximizes system throughput while achieving user-defined efficiency goals for both cost and energy consumption. We implement a small scale prototype of our OSCAR system, which we use to evaluate its performance. Our evaluation shows that we provide up to 150% enhancement in the throughput compared to current operating systems with only minor updates to the client devices.

We present a complete system, called Challenged Content Delivery Network (CCDN), to efficiently deliver multimedia content to mobile users who live in developing countries, rural areas, or over-populated cities with no or weak network infrastructure. These mobile users do not have always-on Internet access. We demo our CCDN, implemented on a Linux server, Raspberry Pi proxies, and Android phones from three aspects: multimedia, networking, and machine learning tools. We propose multiple optimization algorithm modules that compute personalized distribution plans, and maximize the overall user experience. CCDN allows people living in area with challenged networks access to multimedia content, like news reports, using mobile devices, such as smartphones. This in turn will help in eliminating the digital divide, which refers to information inequality to persons with different Internet accessing abilities.

Monitoring breathing rates and patterns helps in the diagnosis and potential avoidance of various health problems. Current solutions for respiratory monitoring, however, are usually invasive and/or limited to medical facilities. In this paper, we propose a novel respiratory monitoring system, UbiBreathe, based on ubiquitous off-the-shelf WiFi-enabled devices. Our experiments show that the received signal strength (RSS) at a WiFi-enabled device held on a person's chest is affected by the breathing process. This effect extends to scenarios when the person is situated on the line-of-sight (LOS) between the access point and the device, even without holding it. UbiBreathe leverages these changes in the WiFi RSS patterns to enable ubiquitous non-invasive respiratory rate estimation, as well as apnea detection.

While WiFi-based indoor localization services are on the rise, existing solutions require periodic updates and therefore exhibit high power demand. In this paper, we propose a novel Social Aware Cluster Based Localization algorithm (SAC-Loc) that leverages social information between nodes in order to dynamically cluster those that exhibit similar mobility patterns. SAC-Loc deploys socially-aware algorithms that dynamically determine when to split and coalesce clusters depending on predicted network topology changes. Based on social ties between encountered nodes, it estimates cluster stability metrics in order to avoid joining crossing nodes with temporary proximity, or the unnecessary splitting of a group due to wireless scanning limitations. We analyze and evaluate our algorithms using a data driven approach based on real-world traces, in addition to an experimental implementation and deployment in our department. While state-of-the-art group localization algorithms can either be energy efficient or highly accurate, SAC-Loc provides a desirable trade-off between accuracy and energy consumption based on popular indoor localization applications.

Mobile Device Clouds are becoming a reality with the quantitative and qualitative upgrades on mobile devices such as smart-phones and tablets. This proliferation renders mobile devices capable of initiating sophisticated cyberattacks especially when they coordinate together and form a distributed mobile botnets which we call “MobiBots”. MobiBots infect a large number of mobile devices and schedule targeted attacks by leveraging device-to-device (d2d) short range wireless communications. In this work, we first introduce MobiBots, the formation of MobiBots, their challenges, and limitations. We then make the case for MobiBots utilization. We show the potential for and impact of the large scale infection and coordination of mobile devices via short range wireless technologies in attacks against other mobile devices that come within proximity. We show that MobiBots are difficult to detect and isolate compared to common botnets. However, prevention techniques cost at least 40% of the network capacity.

Wireless sensor and actuator networks have been extensively deployed for enhancing industrial control processes and supply-chains, and many forms of surveillance and environmental monitoring. The availability of low-cost mobile robots equipped with a variety of sensors in addition to communication and computational capabilities makes them particularly promising in target coverage tasks for ad hoc surveillance, where quick, low-cost or non-lasting visual sensing solutions are required, e.g. in border protection and disaster recovery. In this paper, we consider the problem of low complexity placement and orientation of mobile cameras to cover arbitrary targets. We tackle this problem by clustering proximal targets, while calculating/estimating the camera location/direction for each cluster separately through our cover-set coverage method. Our proposed solutions provide extremely computationally efficient heuristics with only a small increase in number of cameras used, and a small decrease in number of covered targets.

The exponential increase in mobile data demand coupled with the rapid deployment of various wireless access technologies have led to the proliferation of multi-interface enabled devices. As a result, researchers focused on exploiting the available interfaces on such devices in both solitary and coll

Interest and social-awareness can be valuable determinants in decisions related to content delivery in mobile environments. Under certain conditions, we can deliver content with less cost and better delivery ratios, while only involving users that are interested in the type of content being delivered. However, the depletion of valuable power resources poses a deterrent to node participation in such interest-aware forwarding systems. No significant research contribution has been identified to collectively maximize the benefits of social, interest, and power awareness. In this work, we propose a new algorithm called PIPeR which integrates power awareness with an interest and socially aware forwarding algorithm called IPeR. Through simulations, we present and evaluate four modes of PIPeR. The results show that PIPeR is more fair and preserves at least 22% of the power IPeR consumes with less delay, while relying significantly on interested forwarders and with comparable cost to maintain similar delivery ratios.

Cyber-Physical Systems (CPS) have the promise of presenting the next evolution in computing with potential applications that include aerospace, transportation, and various automation systems. These applications motivate advances in the different sub-fields of CPS such as mobile computing, context awareness, and computer vision. However, deploying and testing complete CPSs is known to be a complex and expensive task. In this paper, we present the design, implementation, and evaluation of Up and Away (UnA): a testbed for Cyber-Physical Systems that use Unmanned Aerial Vehicles (UAVs) as their main physical component. UnA aims to abstract the control of physical system components to reduce the complexity of UAV oriented CPS experiments. UnA provides APIs to allow for converting CPS algorithm implementations, developed typically for simulations, into physical experiments using a few simple steps. We present two scenarios of using UnA's API to bring mobile-camera-based surveillance algorithms to life, thus exhibiting the ease of use and flexibility of UnA.

Many mobile applications overcome their device limitations in computational, energy, or data resources by offloading computations to the cloud. In this paper, we consider environments in which computational offloading occurs amongst a set of mobile devices. We call such an environment a mobile device cloud (MDC). In this work, we first highlight the gain in computation time and energy consumption that can be achieved by offloading tasks to nearby devices within an MDC compared to a cloud. We then propose and implement an MDC platform that enables the creation and assessment of various offloading algorithms in MDCs. This platform consists of an Android application deployable across MDC devices, and a test bed to measure power being consumed by a mobile device. We utilize this platform to carry out various offloading experiments on an MDC test bed from which we gain interesting insights into the potential for MDC offloading. Results from these experiments show up to 50% gain in time and 26% gain in energy. Finally, we address the off loadee selection problem in MDCs by proposing several social-based algorithms. The potential promise of this approach is shown by evaluating these algorithms using real data sets that include contact traces and social information of mobile devices in a conference setting.

With the ubiquity of WiFi-enabled smartphones, and large-scale access point deployment, WiFi-based localization is one of the most promising indoor localization systems. Existing WiFi localization solutions, however, exhibit high power demand due to the periodic updates required, which raises the barrier for deployment on mobile devices since battery life is a crucial resource. In this paper, we present an energy efficient architecture, GreenLoc, that leverages the numerous sensors existing on mobile devices, along with typical group mobility patterns, in order to lower the average energy cost for indoor localization. We also propose a cluster-based localization algorithm, integrate it with our GreenLoc architecture, and evaluate its performance via simulation and prototype implementation. Our results show up to 60% reduction in average energy consumed with a small penalty in accuracy that does not impact the performance of the range of applications we target.

Cognitive radio networks (CRNs) provide a solution to increase the utilization of the scarce radio frequency spectrum. Building testbeds for CRNs is one of the main challenges that can affect the wide deployability of such networks. In this paper, we present the design, implementation, and evaluation of CogFrame: a framework that facilitates the development of cost-efficient large-scale CRNs routing protocols testbeds. The framework allows the designers to focus on the CRNs routing protocols by abstracting the PHY and MAC layers while providing the necessary cross layer functionalities. CogFrame works with standard computers and WiFi cards to reduce the cost while allowing integration with other special hardware for more flexibility. In addition, CogFrame provides different modules for implementing and emulating complex scenarios such as regulatory authority policies, mobility management, and topology management. We benchmark the performance of CogFrame and compare it to standard ns-2 simulations and USRP2 implementations. In addition, we case study a location-aided routing protocol for CRNs using both CogFrame and ns-2 simulations. Our results highlight the ease of implementation, low-cost, and realistic replication of the CRN environment, showing the promise of CogFrame as a testbed for future CRNs implementations.

Despite the plethora of opportunistic forwarding solutions offered by the research community, we revisit this domain from a new perspective by exploiting the concept of space syntax to enable deployable solutions in large scale urban environments. We present a set of algorithms that build upon space syntax, which predicts natural movement patterns by interacting with pre-built static environments. We design these algorithms for three assumption categories that represent the spectrum of assumptions regarding the underlying environment and node capabilities. We adopt a data-driven approach to evaluate the performance of our algorithms when compared to other state-of-the-art solutions within each representative category that make similar assumptions. Overall, our results show the great promise space syntax based algorithms have for efficiently guiding messages towards the destination. We show 5% to 20% success rate improvement compared to selected well known state-of-the-art forwarding algorithms within each assumption category while reducing the cost in terms of message replicas by up to 10%.

In this paper, we consider an environment in which computational offoading is adopted amongst mobile devices. We call such an environment a mobile device cloud (MDC). In this work, we highlight via emulation, experimenation and real measurements, the potential gain in computation time and energy consumption that can be achieved by offoading tasks within an MDC. We also propose and develop an experimental platform to enable researchers create and experiment with novel offoading algorithms in MDCs.

Despite extensive research on WiFi indoor localization, very few solutions are widely deployed, largely due to their high energy consumption. In this paper, we propose several energy saving strategies with varying localization accuracy and energy consumption tradeoffs in WiFi indoor localization. Instead of localizing every single device, these strategies exploit short range low-power communication technologies, to localize clusters of mobile devices, via a representative cluster head. We propose various cluster head selection algorithms that offer different trade offs between localization accuracy and power consumption. The outcome of this work provides insights into the effectiveness and cost of a particular strategy depending on the needs of the application requiring varying localization service levels.

Various emerging context aware social-based applications and services assume constant non-disruptive connectivity. Mobile advertisers in such environments want to reach potentially interested users in a given proximity and within a specified short-duration, whether these users are connected to the network or not. While opportunistic forwarding algorithms can be leveraged for forwarding these advertisements, there is little incentive for those not interested in the ad to act as forwarders. Our goal in this paper is to leverage explicit interest, gathered from a user's social profile, and integrate it with social-based opportunistic forwarding algorithms in order to enable soft realtime opportunistic ad delivery in intermittently connected mobile networks. We propose IPeR, a fully distributed interest-aware forwarding algorithm that integrates with PeopleRank to reduce the overall cost and delay while reducing the number of contacted uninterested candidates. Our results, obtained via simulations and validated with real mobility traces coupled with user social data, are promising. In comparison to interest-oblivious socially-aware protocols such as PeopleRank, the IPeR approach reduces the cost to 70% to reach the same delivery ratio, and reduces the ratio of contacted uninterested forwarders by 23%. It also achieves an extra 70% recall and 107% accuracy with only 2% less precision.

The widespread deployment of varying networking technologies, coupled with the exponential increase in end-user data demand, have led to the proliferation of multi-homed or multi-interface enabled devices. To date, these interfaces are mainly utilized one at a time based on network availability, cost, and user-choice. While researchers have focused on simultaneously leveraging these interfaces by aggregating their bandwidths, these solutions however, have faced a steep deployment barrier. In this demo, we demonstrate the OPERETTA optimal bandwidth aggregation system. OPERETTA aims to utilize all the available network interfaces on a mobile device by distributing the users traffic on them. It is implemented as a Layered Service Provider (LSP) in the Windows OS to intercept the network connections from the applications and schedule these connections on the most appropriate interfaces. The network interfaces are chosen with the goal to optimize the overall system power consumption while meeting a certain throughput constraint calculated based on a user defined utility. The demo shows the effect of applying different scheduling algorithms and user utility functions on the system throughput and energy consumption.

The widespread deployment of varying networking technologies, coupled with the exponential increase in end-user data demand, have led to the proliferation of multi-homed or multi-interface enabled devices. To date, these interfaces are mainly utilized one at a time based on network availability, cost, and user-choice. While researchers have focused on simultaneously leveraging these interfaces by aggregating their bandwidths, these solutions however, have faced a steep deployment barrier. In this paper, we propose a novel optimal, energy-efficient, and deployable bandwidth aggregation system (OPERETTA) for multiple interface enabled devices. OPERETTA satisfies three goals: achieving a user defined throughput level with optimal energy consumption over multiple interfaces, deployability without changes to current legacy servers, and leveraging incremental deployment to achieve increased performance gains. We present the OPERETTA architecture and formulate the optimal scheduling problem as a mixed integer programming problem yielding an efficient solution. We evaluate OPERETTA via implementation on the the Windows OS, and further verify our results with simulations on NS2. Our evaluation shows the tradeoffs between the energy and throughput goals. Furthermore, with no modifications to current legacy servers, OPERETTA achieves throughput gains up to 150% compared to current operating systems with the same energy consumption. In addition, with as few as 25% of the servers becoming OPERETTA enabled, OPERETTA performance reaches the throughput upper bound, highlighting its incremental deployment and performance gains.

The widespread deployment of varying networking technologies, coupled with the exponential increase in end-user data demand, have all led to the proliferation of multi-homed or multi-interface enabled devices. To date, these interfaces are mainly utilized one at a time based on network availability, cost, and user-choice. Researchers have recently focused on leveraging these interfaces simultaneously by proposing solutions to aggregate their bandwidths in order to ultimately increase throughput and satisfy the end-user's growing demand on data. These solutions, however, have faced a steep deployment barrier due to various system design choices and heavy demand on energy. In this paper, we propose a novel Green and Deployable Bandwidth Aggregation System (G-DBAS) for multiple interface enabled devices. G-DBAS addresses a set of challenges including automatically estimating the characteristics of applications and scheduling various connections to different interfaces along with meeting different energy consumption goals set by users. We fully implement G-DBAS on the Windows OS and evaluate various scheduling strategies that we propose. Our implementation and simulation results show that G-DBAS can achieve the user energy-throughput goals while operating as an out-of-the-box standard Windows executable, highlighting its deployability and ease of use.

Social forwarding, recently a hot topic in mobile opportunistic networking, faces extreme challenges from potentially large numbers of mobile nodes, vast areas, and limited communication resources. Such conditions render forwarding more challenging in large-scale networks. We observe that forwarding techniques based on social popularity fail to efficiently forward messages in large scale networks. The social popularity of nodes might not scale with the network size in a way that necessarily correlates with the contact opportunities and mobility patterns of these nodes. In this paper, we demonstrate, based on real mobility traces, the weakness of existing social forwarding algorithms in large scale communities. We address this weakness by proposing strategies for partitioning these large scale communities into sub-communities based on geographic locality or social interests. We also examine exploiting particular nodes, named MultiHomed nodes, in order to disseminate messages across these sub- communities. Finally, we introduce CAF, a Community Aware Forwarding framework, which can easily be integrated with the state-of-the-art social forwarding algorithms in order to improve their performance in large scale networks. We use real mobility traces to evaluate our proposed techniques. Our results empirically show a performance increase of around 40% and 5% to 30% better success delivery rates compared to state-of- the-art social forwarding algorithms, while incurring a marginal increase in cost.

The fundamental challenge in opportunistic networking, regardless of the application, is enabling node cooperation in order to forward a message. While node cooperation is considered as a fundamental property in such networks, ensuring such a property between two devices in mobile opportunistic networks remains largely unexplored. In this paper, we investigate the potential impact of the lack of trust on node cooperation. We adopt a real-trace driven approach to study and analyze the trade-off between trust and success delivery rates in opportunistic networks. We first explore leveraging social information to establish trustworthy communication for mobile opportunistic networks. We then propose six trust based filters that use three social-based estimators of trust including common interests, common friends, and the distance in the social graph, coupled with two major techniques of trust establishment including Relay-to-Relay, and Source-to-Relay. We finally show that our trust filters yield a fair trade-off between trust and success rate by achieving more than 35% success rate compared to an untrusted environment where 10% of the nodes refuse to cooperate in the absence of trust.

In the last decade or so, the ACM, the IEEE and other organizations have acknowledged that there is a problem with the way communication is taught in the Computer Science curriculum: the writing, speaking, and presentation skills students learn in the classroom do not match what is expected of them in the workplace. The proposed solution, adopted by many undergraduate colleges, was to add a technical communication course to the CS curriculum. This does not appear to be enough, as mainstream accreditation boards are still emphasizing the need for improvement of communication skills instruction in their recent reports and recommendations. For the last two years, we have experimented with a complementary transversal approach where many "traditional CS" courses in our program have added a communication component to their syllabus, while at the same time our technical communication course has been revamped to expose students to realistic practices as promoted by situated learning theory. The results, so far anecdotal, point to improved student performance and attitude across several communication dimensions, in particular writing and presentation. We plan to develop this experiment by spreading it across more classes and by starting to collect rigorous measurements of students' communication performance.

The fundamental challenge in opportunistic networking, regardless of the application, is when and how to forward a message. Rank-based forwarding techniques currently represent one of the most promising methods for addressing this message forwarding challenge. While these techniques have demonstrated great efficiency in performance, they do not address the rising concern of fairness amongst various nodes in the network. Higher ranked nodes typically carry the largest burden in delivering messages, which creates a high potential of dissatisfaction amongst them. In this paper, we adopt a real-trace driven approach to study and analyze the tradeoff between the efficiency and fairness of rank-based forwarding techniques in mobile opportunistic networks. Our work comprises three major contributions. First, we quantitatively analyze the tradeoffs between fair and efficient environments. Second, we demonstrate how fairness coupled with efficiency can be achieved based on real mobility traces. Third, we propose FOG, a real-time distributed framework to ensure efficiency-fairness tradeoff using local information. Our data-driven experiment and analysis show that mobile opportunistic communication between users may fail with the absence of fairness in participating high-ranked nodes, and an absolute fair treatment of all users yields inefficient communication performance. Finally our analysis show that FOG ensures relative equality in the distribution of resource usage among neighbor nodes while keeping the success rate and cost performance near optimal.

In a Multi-hop Wireless Networks (MHWN), packets are routed between source and destination using a chain of intermediate nodes; chains are a fundamental communication structure in MHWNs whose behavior must be understood to enable building effective protocols. The behavior of chains is determined by a number of complex and interdependent processes that arise as the sources of different chain hops compete to transmit their packets on the shared medium. In this paper, we show that MAC level interactions play the primary role in determining the behavior of chains. We evaluate the types of chains that occur based on the MAC interactions between different links using realistic propagation and packet forwarding models. We discover that the presence of destructive interactions, due to different forms of hidden terminals, does not impact the throughput of an isolated chain significantly. However, due to the increased number of retransmissions required, the amount of band-width consumed is significantly higher in chains exhibiting destructive interactions, substantially influencing the over-all network performance. These results are validated by testbed experiments. We finally study how different types of chains interfere with each other and discover that well behaved chains in terms of self-interference are more resilient to interference from other chains.

Chains or multi-hop paths are the fundamental communication structure in multi-hop wireless networks. Understanding chain behavior is critical in order to build effective higher layer protocols. This paper examines the problem of how MAC level interactions influence chain behavior in a general multi-hop wireless network where multiple chains coexist. We first classify chains based on the MAC interactions observed between its hops when there is no external traffic. Then we identify the interactions across two interfering chains for the most common categories of chains. We study the probability of occurrence, and estimate the effect of MAC interactions on the performance of the chains. We also show that different chains exhibit different transmission patterns; this is an effect that is necessary for accurately estimating chain performance. We observe that destructive interactions arise more frequently among two interfering chains than they do within a single chain. Moreover, chains that have hidden terminals due to self-interference are more prone to have cross-chain hidden terminals. Thus, both intra-chain as well as cross-chain interactions, ultimately provide significant insight into how chains interact.

Abstract The incredible growth in the capabilities and functionality of mobile devices has enabled new applications and network architectures to emerge. Due to the potential for node mobility, along with significant node heterogeneity, characteristics such as very large delays, intermittent links, and high link error rates pose a new set of network challenges. Along with these challenges, end‐to‐end paths are assumed not to exist and message relay approaches are often adopted. While message flooding is a simple and robust solution for such cases, its cost in terms of network resource consumption is unaffordable. In this paper, we focus on the evaluation of different controlled message flooding schemes over disconnected sparse mobile networks. We study the effect of these schemes on message delay, network resource consumption, and neighbor discovery overhead. Our simulations show that our schemes can save substantial network resources while incurring a negligible increase in the message delivery delay. Copyright © 2008 John Wiley & Sons, Ltd.

The rapid increase and proliferation of mobile wireless technologies has led to the rise of new problems that fall under the umbrella of challenged networks, namely, intermittent network connectivity. This diversity in wireless devices, along with their convergence, has granted users access to multiple heterogeneous networks available in parallel. Nowadays, a user generally expects to be connected in all places at all times. To better meet this expectation, we propose a system that takes advantage of intermittent connection opportunities while exploiting other networks available in parallel. We build our system over the parallel networks architecture, ParaNets, which suggests using heterogeneous networks simultaneously as data and control channels. Our system adopts the Data Bundling System for Intermittent Connections (DBS-IC), previously proposed as a stand-alone architecture for intermittent connectivity, and integrates it with the ParaNets architecture. We evaluate our system by fully implementing a ParaNets-enabled version of DBS-IC and thoroughly testing it over emulated network conditions. Characteristics of these networks are adopted based on real-life data on 802.11, 3G cell, and satellite networks. Our results show how minimal exploitation of parallel networks largely optimizes both cost and delivery rate.

As mobile and wireless technologies become more pervasive in our society, people begin to depend on network connectivity regardless of their location. Their mobility, however, implies a dynamic topology where routes to a destination cannot always be guaranteed. The intermittent connectivity, which results from this lack of end-to-end connection, is a dominant problem that leads to user frustration. Existing research to provide the mobile user with a mirage of constant connectivity generally presents mechanisms to handle disconnections when they occur. In contrast, the system we propose in this paper provides ways to handle disconnections before they occur. We present a data bundling system for intermittent connections (DBS-IC) comprised of a stationary agent (SA) and a mobile agent (MA). The SA proactively gathers data the user has previously specified, and opportunistically sends this data to the MA. The SA groups the user-requested data into one or more data bundles, which are then incrementally delivered to the MA during short periods of connectivity. We fully implement DBS-IC and evaluate its performance via live tests under varying network conditions. Results show that our system decreases data retrieval time by a factor of two in the average case and by a factor of 20 in the best case

The evolution of wireless devices along with the increase in user mobility have created new challenges such as network partitioning and intermittent connectivity. These new challenges have become apparent in many situations where the transmission of critical data is of high priority. Disaster rescue groups, for example, are equipped with numerous devices which constantly gather and transmit various forms of data. The challenge of establishing communication between groups of this type has led to an evolutionary form of networks which we consider in this paper, namely, delay tolerant mobile networks (DTMNs). Nodes in DTMNs usually form clusters that we define as regions. Nodes within each region have end-to-end paths between them. Both regions, as well as nodes within a region, can be either stationary or mobile. For such environments, we propose using a dedicated set of messengers that relay message bundles between these regions. Our goal is to understand how messenger scheduling can be used to improve network performance and connectedness. We develop several classes of messenger scheduling algorithms which can be used to achieve inter-regional communication in such environments. We use simulation to better understand the performance and tradeoffs between these algorithms