This paper describes Mercury, a wearable, wireless sensor platform for motion analysis of patients being treated for neuromotor disorders, such as Parkinson’s Disease, epilepsy, and stroke. In contrast to previous systems intended for short-term use in a laboratory, Mercury is designed to support long-term, longitudinal data collection on patients in hospital and home settings. Patients wear up to 8 wireless nodes equipped with sensors for monitoring movement and physiological conditions. Individual nodes compute high-level features from the raw signals, and a base station performs data collection and tunes sensor node parameters based on energy availability, radio link quality, and application specific policies. Mercury is designed to overcome the core challenges of long battery lifetime and high data fidelity for long-term studies where patients wear sensors continuously 12 to 18 hours a day. This requires tuning sensor operation and data transfers based on energy consumption of each node and processing data under severe computational constraints. Mercury provides a high-level programming interface that allows a clinical researcher to rapidly build up different policies for driving data collection and tuning sensor lifetime. We present the Mercury architecture and a detailed evaluation of two applications of the system for monitoring patients with Parkinson’s Disease and epilepsy.

Abstract—We describe a wireless wearable system that was developed to provide quantitative gait analysis outside the confines of the traditional motion laboratory. The sensor suite includes three orthogonal accelerometers, three orthogonal gyroscopes, four force sensors, two bidirectional bend sensors, two dynamic pressure sensors, as well as electric field height sensors. The “GaitShoe ” was built to be worn in any shoe, without interfering with gait and was designed to collect data unobtrusively, in any environment, and over long periods. The calibrated sensor outputs were analyzed and validated with results obtained simultaneously from the Massachusetts General Hospital, Biomotion Laboratory. The GaitShoe proved highly capable of detecting heel-strike and toe-off, as well as estimating foot orientation and position, inter alia. Index Terms—Biomedical measurements, body sensor networks, legged locomotion, multisensor systems, telemetry. I.

In this paper, we consider the challenge of preserving patient privacy in the context of mobile healthcare and home-care systems, that is, the use of mobile computing and communications technologies in the delivery of healthcare or the provision of at-home medical care and assisted living. This paper makes three primary contributions. First, we compare existing privacy frameworks, identifying key differences and shortcomings. Second, we identify a privacy framework for mobile healthcare and home-care systems. Third, we extract a set of privacy properties intended for use by those who design systems and applications for mobile healthcare and home-care systems, linking them back to the privacy principles. Finally, we list several important research questions that the community should address. We hope that the privacy framework in this paper can help to guide the researchers and developers in this community, and that the privacy properties provide a concrete foundation for privacysensitive systems and applications for mobile healthcare and home-care systems.

We propose a distributed recognition framework to classify continuous human actions using a low-bandwidth wearable motion sensor network, called distributed sparsity classifier (DSC). The algorithm classifies human actions using a set of training motion sequences as prior examples. It is also capable of rejecting outlying actions that are not in the training categories. The classification is operated in a distributed fashion on individual sensor nodes and a base station computer. We model the distribution of multiple action classes as a mixture subspace model, one subspace for each action class. Given a new test sample, we seek the sparsest linear representation of the sample w.r.t. all training examples. We show that the dominant coefficients in the representation only correspond to the action class of the test sample, and hence its membership is encoded in the sparse representation. Fast linear solvers are provided to compute such representation via ℓ 1-minimization. To validate the accuracy of the framework, a public wearable action recognition database is constructed, called wearable action recognition database (WARD). The database is comprised of 20 human subjects in 13 action categories. Using up to five motion sensors in the WARD database, DSC achieves state-of-the-art performance. We further show that the recognition precision only decreases gracefully using smaller subsets of active sensors. It validates the robustness of the distributed recognition framework on an unreliable wireless network. It also demonstrates the ability of DSC to conserve sensor energy for communication while preserve accurate global classification.

We propose a distributed recognition method to classify human actions using a low-bandwidth wearable motion sensor network. Given a set of pre-segmented motion sequences as training examples, the algorithm simultaneously segments and classifies human actions, and it also rejects outlying actions that are not in the training set. The classification is distributedly operated on individual sensor nodes and a base station computer. We show that the distribution of multiple action classes satisfies a mixture subspace model, one subspace for each action class. Given a new test sample, we seek the sparsest linear representation of the sample w.r.t. all training examples. We show that the dominant coefficients in the representation only correspond to the action class of the test sample, and hence its membership is encoded in the representation. We further provide fast linear solvers to compute such representation via ℓ 1-minimization. Using up to eight body sensors, the algorithm achieves state-of-the-art 98.8% accuracy on a set of 12 action categories. We further demonstrate that the recognition precision only decreases gracefully using smaller subsets of sensors, which validates the robustness of the distributed framework. 1.

Abstract—Networked mobile devices have great potential to enable individuals (and their physicians) to better monitor their health and to manage medical conditions. In this paper, we examine the privacy-related threats to these so-called mHealth technologies. We develop a taxonomy of the privacy-related threats, and discuss some of the technologies that could support privacy-sensitive mHealth systems. We conclude with a brief summary of research challenges. I.

This paper introduces a compact, wireless, wearable system that measures signals indicative of forces, torques and other descriptive and evaluative features that the human body undergoes during bursts of extreme physical activity (such as during athletic performance). Standard approaches leverage highspeed camera systems, which need significant infrastructure and provide limited update rates and dynamic accuracy. This project uses 6 degree-offreedom inertial measurement units worn on various segments of an athlete’s body to directly make these dynamic measurements. A combination of low and high range sensors enables sensitivity for both slow and fast motion, and the addition of a compass helps in tracking joint angles. Data from the battery-powered nodes is acquired using a custom wireless protocol over an RF link and analyzed offline. Several professional pitchers and batters were instrumented with the system and data was gathered over many pitches and swings. We show some biomechanically descriptive parameters extracted from this data, and highlight ongoing work and system improvements. 1. Introduction & Prior

Abstract—To maximize the communication throughput for wireless sensing systems, designers have attempted various combinations of protocol design and manual code optimization. Although the theoretical bandwidth limit is easy to determine loosely, there have been no systematic ways to arrive at a tight upper-bound. One contribution of this paper is a formula for deriving a tight upper-bound on the throughput of low-complexity wireless interfaces transmitting packets of a fixed size. It takes into account not only the software execution times on the nodes but also other communication protocols that must be bridged by the base station. The proposed upper-bound, which we believe is the tightest, represents the maximum amount of bandwidth utilization that can be achieved in practice. It can also serve as a means of comparing protocols built on different platforms. Another contribution is a streamlined schedule-based protocol, called RIPE-MAC, which achieves at least 83 % of the upper-bound, significantly higher than previously achieved throughput. The proposed protocol needs no clock synchronization and incurs no further complexity on sensor nodes. In the proposed protocol, synchronization and schedule updates are reduced to a single pull message. Keywords—wireless sensor networks, medium access control protocols, communication scheduling, high-data-rate monitoring, analytical upper-bound on throughput, time division multiple access I.

In this thesis I demonstrate a framework for a wireless sensor-based mobile music environment. Most prior work has not been truly portable. Those that were have focused on external data as opposed to properties of the listener. In this project I built a short-range wireless sensor network (using the ZigBee protocol and an accelerometer) and a compiler for PureData, a graphical music processing language. With these parts, I realized a synchronized music experience that generates a soundtrack based on the listener’s movement. By synchronizing the music to the user’s natural rhythms, it encourages the user to maintain a given pace for a longer period of time. I describe extensions to this example that point to a future of portable interactive music tied to exercise and physical activity.

Abstract As microelectronics have escalated in capability via Moore’s Law, electronic sensors have similarly advanced. Rather than dedicate a small number of sensors to hardwired designs that expressly measure parameters of interest, we can begin to envision a near future with sensors as commodity where dense, multimodal sensing is the rule rather than the exception, and where features relevant to many applications are dynamically extracted from a rich data stream. This article surveys a series of projects at the MIT Media Lab’s Responsive Environments Group that explore various embodiments of such agile sensing structures, including high-bandwidth, wireless multimodal sensor clusters, massively distributed, ultra-low-power "featherweight " sensor nodes, and extremely dense sensor networks as digital "skins". This paper also touches on other examples involving gesture sensing for large interactive surfaces and interactive media, plus overviews projects in parasitic power harvesting.