Search

Five cellphones are sold every second, and there are four times more cellphones than computers, meaning there are some billions of mobile handheld devices in existence. Modern cellphones are equipped with multiple context sensors used by increasingly sophisticated software applications that exploit the sensors, allowing the applications to adapt automatically to changes in the surrounding environment, such as by responding to the location and speed of the user. The architecture of such applications is typically layered and incorporates a context-awareness middleware to support processing of context values. While this layered architecture is very natural for the design and implementation of applications, it gives rise to new kinds of faults and faulty behavior modes, which are difficult to detect using existing validation techniques. In this paper we provide scenarios illustrating such faults and exploring how they manifest in context-aware adaptive applications. [ABSTRACT FROM AUTHOR]

Numerous studies have been conducted to improve the features of the "smart" canes. However, there is insufficient research performed to develop the guidelines to create a usable smart cane. In this paper, a Dynamic Sensor Orientation Unit (DSOU) is introduced to maintain a detection distance irrespective of the angle of the cane. The paper also aims to outline guidelines for placement of the Obstacle Detection System (ODS) and the sensor angle to be used in our Intelligent Mobility Cane (IMC) prototypes. A usability test with 11 people who use a white cane for daily navigation was conducted to study the users' preference and behavior with a fixed sensor angle IMC. The analysis revealed the angle of the sensor is dependent on the participant's height and how users hold the canes. The results suggest requirements for sensor angle orientation and outline guidelines for ODS designs on future IMCs. [ABSTRACT FROM AUTHOR]

Crown-of-Thorn Starfish (COTS) outbreaks are a major cause of coral loss on the Great Barrier Reef (GBR) and substantial surveillance and control programs are ongoing to manage COTS populations to ecologically sustainable levels. In this paper, we present a comprehensive real-time machine learning-based underwater data collection and curation system on edge devices for COTS monitoring. In particular, we leverage the power of deep learning-based object detection techniques, and propose a resource-efficient COTS detector that performs detection inferences on the edge device to assist marine experts with COTS identification during the data collection phase. The preliminary results show that several strategies for improving computational efficiency (e.g., batchwise processing, frame skipping, model input size) can be combined to run the proposed detection model on edge hardware with low resource consumption and low information loss. [ABSTRACT FROM AUTHOR]

In a near future it is expected that most things we rely on in our everyday life will contain sensors and electronic based information, have computing power, run embedded software and connect to networks. A multitude of heterogeneous things will operate in a highly dynamic environment and will collaborate with other connected systems and things to provide users with adaptable services. Constructing and controlling such adaptive things is complex. A main challenge is to cope with the dynamicity which requires the things to autonomously adapt to various execution contexts. In this paper we present an approach to develop adaptive firmwares for devices which do not have the resources to rely on advanced operating systems, middlewares or frameworks to support runtime adaptation. The paper is illustrated with the example of an adaptive temperature sensor network running on a microcontroller platform. [ABSTRACT FROM AUTHOR]

Recent advances in deep neural networks (DNNs) have shown Binary Neural Networks (BNNs) are able to provide a reasonable accuracy on various image datasets with a significant reduction in computation and memory cost. In this paper, we explore two BNNs: hybrid BNN (HBNN) and XNORBNN, where the weights are binarized to +1/-1 while the neuron activations are binarized to 1/0 and +1/-1, respectively. Two SRAM bit cell designs are proposed, namely, 6T SRAM for HBNN and customized 8T SRAM for XNOR-BNN. In our design, the high-precision multiply-and-accumulate (MAC) is replaced by bitwise multiplication for HBNN or XNOR for XNOR-BNN plus bit-counting operations. To parallelize the weighted sum operation, we activate multiple word lines in the SRAM array simultaneously and digitize the analog voltage developed along the bit line by a multi-level sense amplifier (MLSA). In order to partition the large matrices in DNNs, we investigate the impact of sensing bit-levels of MLSA on the accuracy degradation for different sub-array sizes and propose using the nonlinear quantization technique to mitigate the accuracy degradation. With 64x64 sub-array size and 3-bit MLSA, HBNN and XNORBNN architectures can minimize the accuracy degradation to 2.37% and 0.88%, respectively, for an inspired VGG-16 network on the CIFAR-10 dataset. Design space exploration of SRAM based synaptic architectures with the conventional row-by-row access scheme and our proposed parallel access scheme are also performed, showing significant benefits in the area, latency and energy-efficiency. Finally, we have successfully taped-out and validated the proposed HBNN and XNOR-BNN designs in TSMC 65 nm process with measured silicon data, achieving energyefficiency >100 TOPS/W for HBNN and >50 TOPS/W for XNOR-BNN. [ABSTRACT FROM AUTHOR]

The past decade's research in visible light positioning (VLP) has led to technologies with high location precision. However, existing VLP systems either require specialized LEDs which hinder large-scale deployment, or need cameras which preclude continuous localization because of high power consumption and short coverage. In this paper, we propose Pulsar, which uses a compact photodiode sensor, readily fit into a mobile device, to discriminate existing ceiling lights--either fluorescent lights or conventional LEDs--based on their intrinsic optical emission features. To overcome the photodiode's lack of spatial resolution, we design a novel sparse photogrammetry mechanism, which resolves the light source's angle-of-arrival, and triangulates the device's 3D location and even orientation. To facilitate ubiquitous deployment, we further develop a light registration mechanism that automatically registers the ceiling lights' locations as landmarks on a building's floor map. Our experiments demonstrate that Pulsar can reliably achieve decimeter precision in both controlled environment and large-scale buildings. [ABSTRACT FROM AUTHOR]

Wireless sensors have enabled a number of key applications. Due to their energy constraints, wireless sensors today communicate occasional short samples or pre-determined summary statistics of the data they collect. This means that computing every additional statistic at high fidelity incurs additional communication and energy overhead. This paper presents Joltik, a framework enabling general, future-proof, and energy-efficient analytics for low power wireless sensors. Joltik is general in that it summarizes sensed data from low-power devices without making assumptions on which specific statistical metric(s) are desired at the cloud and is future-proof, meaning it supports new, unforeseen metrics. Joltik is built upon recent theoretical advances in universal sketching, which can enable a Joltik sensor node to report a compact summary of observed data to enable a large class of statistical summaries. We address key system design and implementation challenges with respect to communication, memory, and computation bottlenecks that arise in practically realizing the potential benefits of universal sketching in the low-power regime. We present a proof-of-concept testbed evaluation of Joltik in LoRaWAN NUCLEO-L476RG boards and sensors. Across a range of realistic datasets, Joltik provides up to a 24.6× reduction in energy cost compared to transmitting raw data and outperforms many natural alternatives (e.g., sub-sampling, custom sketches, compressed sensing, and lossy compression) in terms of energy-accuracy trade-offs. [ABSTRACT FROM AUTHOR]

Reliability is a major concern in chip multi-processors (CMPs) due to shrinking technology and low operating voltages. Today's processors designed at sub-10nm technology nodes have high device densities and fast switching frequencies that cause fluctuations in supply voltage (Vdd) and ground networks, which can adversely affect the execution of applications running on them. In this paper, we propose a novel runtime framework to reduce the power supply noise (PSN) in cores and routers at runtime. Experimental results for 7nm FinFET process nodes show that our framework not only achieves up to 4.5x reduction in PSN, and up to 34.3% improvement in application performance, but also manages to map up to 38% more applications when the CMP is oversubscribed, compared to the state-of-the-art. [ABSTRACT FROM AUTHOR]

Limited processing power and memory prevent realization of state of the art algorithms on the edge level. Offloading computations to the cloud comes with tradeoffs as compression techniques employed to conserve transmission bandwidth and energy adversely impact accuracy of the algorithm. In this paper, we propose collaborative processing to actively guide the output of the sensor to improve performance on the end application. We apply this methodology to smart surveillance specifically the task of object detection from video. Perceptual quality and object detection performance is characterized and improved under a variety of channel conditions. [ABSTRACT FROM AUTHOR]

In this paper, we present a low-maintenance, wind-powered, batteryfree, biocompatible, tree wearable, and intelligent sensing system, namely IoTree, to monitor water and nutrient levels inside a living tree. IoTree system includes tiny-size, biocompatible, and implantable sensors that continuously measure the impedance variations inside the living tree's xylem, where water and nutrients are transported from the root to the upper parts. The collected data are then compressed and transmitted to a base station located at up to 1.8 kilometers (approximately 1.1 miles) away. The entire IoTree system is powered by wind energy and controlled by an adaptive computing technique called block-based intermittent computing, ensuring the forward progress and data consistency under intermittent power and allowing the firmware to execute with the most optimal memory and energy usage. We prototype IoTree that opportunistically performs sensing, data compression, and long-range communication tasks without batteries. During in-lab experiments, IoTree also obtains the accuracy of 91.08% and 90.51% in measuring 10 levels of nutrients, NH3 and K2O, respectively. While tested with Burkwood Viburnum and White Bird trees in the indoor environment, IoTree data strongly correlated with multiple watering and fertilizing events. We also deployed IoTree on a grapevine farm for 30 days, and the system is able to provide sufficient measurements every day. [ABSTRACT FROM AUTHOR]

The term "Smart Home" refers to a home that has both a set of sensors to observe the environment and a set of actuators to automatically control home devices to improve the occupants' experience. A Smart Home has the potential to provide a variety of services based on information gleaned by data recorded by a multiplicity of sensor types, but also requires a systematic means (a) of abstracting differences among sensor types and (b) of supporting services and their rapid development. In this paper, we describe our approach towards meeting those needs, including a component of sensor adapters and a language for modeling person activities that can guide the development of health-care services. [ABSTRACT FROM AUTHOR]

One of the important factors in ensuring the correct functionality of wireless sensor networks (WSNs) is the reliability of the software running on individual sensor nodes. Research has shown that path-sensitive static analysis is effective for bug detection and fault diagnosis; however, path-sensitive analysis is prohibitively expensive when applied to a WSN application due to the large state space caused by arbitrary interrupt preemptions. In this paper, we propose a new execution model called lazy preemption that reduces this state space by restricting interrupt handlers to a set of predetermined preemption points, if possible. This execution model allows us to represent the program with an inter-interrupt control flow graph (IICFG), which is easier to analyze than the original CFGs with arbitrary interrupt preemptions. [ABSTRACT FROM AUTHOR]

While the health benefits of exercise are wide-ranging and wellknown, the population of the United States is suffering from a lack of physical activity. We believe that combining elements of social interaction with exercise in video games will lead to increased and sustained engagement in physical activity. In this paper, we present an initial design of GameChanger, a middleware to support the development of a new generation of social exergames that interweave physical activity as a core game mechanic with social elements such as competition and collaboration. The GameChanger middleware provides programming abstractions that are specific to social exergame mechanics, elevating their description so that even non-expert programmers can create interesting social exergames that utilize mobile phones and sensing technology to integrate physical activity into gameplay. In addition, the middleware provides constructs for performing continuous assessment of physical activity during gameplay; these constructs can be used to provide feedback to the gameplayer or to collect datasets for evaluation by health researchers. As such, the GameChanger middleware can also serve as a platform to support scientific experimentation and exploration to determine which combinations of social and physical elements have the greatest impact on physical activity. [ABSTRACT FROM AUTHOR]

In this paper, we analyze and describe the achievable performance of a Wireless Sensor Network (WSN) testbed platform, which has been deployed at the Center of Excellence DEWS (University of L'Aquila, Italy) for positioning applications, and which is implemented using CC2431 sensor nodes developed by Texas Instruments® (TI)/Chipcon. In particular, we will report on an extensive campaign of measurements aiming at analyzing the performance of the CC2431's hardware location-finder engine when localization needs to be performed in a realistic and dynamic indoor propagation environment. Experimental results will show that the need of site-specific parameters for the correct operation of the location engine may severely reduce the localization accuracy of the system in dynamic environments. A simple solution to counteract this problem will also be introduced, and the related improvement will be quantified via experiments. [ABSTRACT FROM AUTHOR]

Off-chip serial buses are common in embedded systems, and due to the long physical lines, can contribute significantly to their energy consumption. However, these buses are often connected to analog sensors, whose data is inherently affected by noise and A/D errors. Thus, communication can tolerate small approximations, without a significant impact on the system outputs quality. In this paper we propose an energy-efficient approximate serial encoding for sensors data, inspired by the T0 technique for parallel buses. Despite its small encoder/decoder overheads, this method is capable of significantly reducing dynamic power (> 90%) in off-chip serial lines, with negligible average error (<1%) on decoded data. [ABSTRACT FROM AUTHOR]

This paper finds that actions of the teeth, namely tapping and sliding, produce vibrations in the jaw and skull. These vibrations are strong enough to propagate to the edge of the face and produce vibratory signals at an earphone. By re-tasking the earphone speaker as an input transducer - a software modification in the sound card - we are able to sense teeth-related gestures across various models of ear/headphones. In fact, by analyzing the signals at the two earphones, we show the feasibility of also localizing teeth gestures, resulting in a human-to-machine interface. Challenges range from coping with weak signals, distortions due to different teeth compositions, lack of timing resolution, spectral dispersion, etc. We address these problems with a sequence of sensing techniques, resulting in the ability to detect 6 distinct gestures in real-time. Results from 18 volunteers exhibit robustness, even though our system - EarSense - does not depend on per-user training. Importantly, EarSense also remains robust in the presence of concurrent user activities, like walking, nodding, cooking and cycling. Our ongoing work is focused on detecting teeth gestures even while music is being played in the earphone; once that problem is solved, we believe EarSense could be even more compelling. [ABSTRACT FROM AUTHOR]

This paper presents an approach to verifying the activities of daily living of older adults at their home. We verify activities, instead of inferring them, because our monitoring approach is driven by routines, initially sketched by users in their environment. Monitoring is supported by a lightweight sensor infrastructure, comprising non-intrusive, low-cost, wireless devices. Verification is performed by applying a simple formula to sensor log data, for each activity of interest. The result value determines whether an activity has been performed. We have conducted an experimental study to validate our approach. To do so, four participants have been monitored during five days at their home, equipped with sensors. When applied to the log data, our formulas were able to automatically verify that a list of activities were performed. They produced the same interpretations, using Signal Detection Theory, as a third party, manually analyzing the log data. [ABSTRACT FROM AUTHOR]

The lack of floor plans is a critical reason behind the current sporadic availability of indoor localization service. Service providers have to go through effort-intensive and time-consuming business negotiations with building operators, or hire dedicated personnel to gather such data. In this paper, we propose Jigsaw, a floor plan reconstruction system that leverages crowdsensed data from mobile users. It extracts the position, size and orientation information of individual landmark objects from images taken by users. It also obtains the spatial relation between adjacent landmark objects from inertial sensor data, then computes the coordinates and orientations of these objects on an initial floor plan. By combining user mobility traces and locations where images are taken, it produces complete floor plans with hallway connectivity, room sizes and shapes. Our experiments on 3 stories of 2 large shopping malls show that the 90-percentile errors of positions and orientations of landmark objects are about 1 ~2m and 5 ~ 9°,while the hallway connectivity is 100% correct. [ABSTRACT FROM AUTHOR]

Traditionally, biomolecular systems have been studied in ensemble. While much can be determined with ensemble measurements, scientific and technological interest is rapidly moving to single-molecule techniques, which rely primarily on fluorescent markers and advanced microscopy techniques. In this paper, we describe recent work using nanoscale transistors based on carbon nanotubes as charge-sensitive detectors. We show carbon nanotubes can be used for ensemble studies through sidewall adsorption. Sensitivity can be greatly enhanced though an engineered defect in the nanotube. Biomolecular interactions are characterized by random-telegraph-noise response, which can be analyzed to study single-molecule kinetics and thermodynamics. [ABSTRACT FROM AUTHOR]

The increased power densities of multi-core processors and the variations within and across workloads lead to runtime thermal hot spots locations of which change across time and space. Thermal hot spots increase leakage, deteriorate timing, and reduce the mean time to failure. To manage runtime thermal variations, circuit designers embed within-die thermal sensors that acquire temperatures at few selected locations. The acquired temperatures are then used to guide runtime thermal management techniques. The capabilities of these techniques are essentially bounded by the spatial thermal resolution of the sensor measurements. In this paper we characterize temperature signals of real processors and demonstrate that on-chip thermal gradients lead to sparse signals in the frequency domain. We exploit this observation to (1) devise thermal sensor allocation techniques, and (2) devise signal reconstruction techniques that fully characterize the thermal status of the processor using the limited number of measurements from the thermal sensors. To verify the accuracy of our methods, we compare our temperature characterization results against thermal measurements acquired from a state-of-the-art infrared camera that captures the mid-band infrared emissions from the back of the die of a 45 nm dual-core processor. Our results show that our techniques are capable of accurately characterizing the temperatures of real processors. [ABSTRACT FROM AUTHOR]

Body Area Networks (BANs) can perform the task of continuous, remote monitoring of a patient's physiological signals in diverse environments. Apart from providing healthcare professionals with extensive logs of a patient's physiological history, BANs can be used to identify and react to emergency situations. We identify three important factors that afflict wireless communication in BANs: impermeability of the human body to radio waves at frequencies commonly used in BANs, efficient operation in mobile and time-varying environments, and mission-critical requirements for quick response to emergencies. An understanding of the link layer behavior of wireless sensor nodes placed on the body is crucial to address these and other challenges such as reducing energy consumption and increasing network lifetime. In this paper, we investigate link layer behavior by placing nodes on the body and directly measuring metrics of interest to engineers such as packet delivery ratio (PDR) and RSSI. Emulating a possible real-life BAN operating at the 2.4 GHz band with 12 sensor nodes, we collect over 80 hours of data from 14 volunteers in 3 different environments that BANs are expected to operate in. We analyze the data to reveal several link layer characteristics to provide insight and guidelines for the designing of BANs. We also evaluate the performance of common routing metrics on our data. Our analysis helps us make the following conclusions. Link PDR is highly affected by the environment and not significantly by the volunteer for the experiment. Routing between nodes on the same side of the body is preferred to routing between nodes on the opposite sides. For links with the same source, failure of packet transmission to a certain node, in some cases, implies the increased probability of reception for other nodes. Most errors occur in bursts of length 1, but a small fraction occur in longer periods (40 packets or more). [ABSTRACT FROM AUTHOR]

Wireless Sensor Networks are emerging as a key sensing technology, with diverse military and civilian applications. In these networks, a large number of sensors perform distributed sensing of a target field. Each sensor is a small battery-operated device that can sense events of interest in its sensing range and can communicate with neighboring sensors. A sensor cover is a subset of the set of all sensors such that every point in the target field is in the interior of the sensing ranges of at least k different sensors in the subset, where k is a given positive integer. The lifetime of the network is the time from the point the network starts operation until the set of all sensors with non-zero remaining energy does not constitute a sensor cover. An important goal in sensor networks is to design a schedule, that is, a sequence of sensor covers to activate in every time slot, so as to maximize the lifetime of the network. In this paper, we design a polynomial-time, distributed algorithm for maximizing the lifetime of the network and prove that its lifetime is at most a factor O(log n * log nB) lower than the maximum possible lifetime, where n is the number of sensors and B is an upper bound on the initial energy of each sensor. Our algorithm does not require knowledge of the locations of nodes or directional information, which is difficult to obtain in sensor networks. Each sensor only needs to know the distances between adjacent nodes in its transmission range and their sensing radii. In every slot, the algorithm first assigns a weight to each node that is exponential in the fraction of its initial energy that has been used up so far. Then, in a distributed manner, it finds a O(log n) approximate minimum weight sensor cover which it activates in the slot. Our simulations reveal that our algorithm substantially outperforms several existing lifetime maximization algorithms. [ABSTRACT FROM AUTHOR]

A SEA Swarm (Sensor Equipped Aquatic Swarm) moves as a group with water current and enables 4D (space and time) monitoring of local underwater events such as contaminants and intruders. For prompt alert reporting, mobile sensors forward events to mobile sinks (i.e., autonomous underwater vehicles) via geographic routing, thus requiring a location service. In this paper, we analyze various design choices to realize an efficient location service in a SEA Swarm. We find that conventional ad hoc network location service protocols cannot be directly used, because the entire swarm moves along water current. We show that maintaining location information in a 2D plane is an optimal design choice. Given this, we propose a bio-inspired location service called a Phero-Trail location service protocol. In Phero-Trail, location information is stored in a 2D upper hull of a SEA Swarm, and a mobile sink uses its trajectory (à la a pheromone trail of ants) projected to the 2D hull to maintain location information. This enables mobile sensors to efficiently locate a mobile sink via an expanding spiral curve search. Our preliminarily results show that Phero-Trail performs better than existing approaches. [ABSTRACT FROM AUTHOR]

Android malware has been found on various third-party online markets, which poses drastic threats to mobile users in terms of security and privacy. Machine learning is one of the promising approaches to discriminate the malicious applications from the benign ones. Despite its higher malware detection capability, a significant challenge remains: in adversarial environment, an attacker can adapt by maximally sabotaging classifiers by polluting training data. This paper proposes KuafuDet, a two-phase learning enhancing approach that adversarially detects the Android malware. Experiments on more than 50,000 Android applications demonstrate the effectiveness and scalability of our approach. [ABSTRACT FROM AUTHOR]

This paper describes the development of a working prototype of a novel free-space haptic human-computer interface called MIDAS Touch. MIDAS Touch works by applying physical forces to a user's finger through the production of a dynamic magnetic field. The magnetic field strength is adjusted in real-time based on the user's movement of his/her finger. A user's hand/finger motion in the real world is mapped to movement of a virtual finger in a virtual world through the use of a Leap Motion 3D tracking sensor. As a person's virtual finger collides with objects in the virtual world, the magnetic field strength is varied. In this demo, we present a case of MIDAS Touch coupled to a standard PC as a computer drawing viewer and drawing application for helping individuals with visual impairment feel what they or others have drawn. [ABSTRACT FROM AUTHOR]

Future networks are ones in which mobile users are immersed in sensor networks. Application domains such as smart homes, intelligent construction sites, and social networking require users moving through an environment to collect and process data available in that environment. However, even as sensors become increasingly prevalent in our environments, only a handful of applications are available to interact with and process the available sensor data. Poor usability accounts for a good portion of this disparity. In this paper we discuss our directions in creating an architecture for making such sensor data usable by application developers. This has the potential to widen the accessibility of developing mobile applications on embedded sensor networks beyond the niche markets it now enjoys. [ABSTRACT FROM AUTHOR]