In-Network, Physical Adaptation of Sensor Networks

Опубликовано 6 сентября 2016, 16:29

Sensor network applications have rapidly evolved as driven by new and urgent demands in environmental monitoring as well as by technology advances that enable new platforms. At the same time, sensor network systems also may take advantage of ubiquitous consumer wireless platforms. It is likely that some of the most interesting progress yet in the sensor network field will occur now as a result of this recent convergence. This presentation will describe recent research results that combine new platforms, applications, and now new opportunities. First, beginning several years ago, a program at UCLA was developed to combine sensing, actuation, and networking, Networked Infomechanical Systems (NIMS). The NIMS vision is the development of new sensor network systems that apply in-field, autonomous, physical reconfigurability to adapt and directly address the uncertainty that confronts measurement. NIMS first accomplishments were in the area of a new actuated sensor system that provides, precise, reliable, energy-efficient, location control in sensor networks. This led to a series of new results in environmental monitoring, in particular for water resource monitoring. The capabilities of this system have expanded and its applications have become yet more urgent. NIMS are currently being applied to characterizing what is considered the most important problem in agricultural water resources in the Western US. The NIMS program has also addressed the requirements for sensor network node systems that apply reconfiguration to optimize platform energy efficiency. This is developed to permit sensor nodes to adapt to uncertain future operation demand. This has led to the development of the Low Power Energy Aware Processing (LEAP) platform. LEAP progress has advanced and now includes two platform classes. Most importantly, LEAP has recently acquired a new capability for high speed energy monitoring that enables per-process and per-component real-time energy profiling integrated with the LEAP platform. This is expected to permit advances in energy-efficiency for embedded computing platforms dedicated to sensing objectives. Finally, the NIMS program, with its focus on in-field reconfigurability, and new node platforms, has also developed applications in the area of biomedical monitoring where consumer electronic devices are integrated with wearable sensor networks in the new field of Telehealth. As for the previous examples, uncertainty appears in the measurement problem and this can also be addressed by new adaptation methods. The research progress in these programs and the corresponding new application opportunities will be described.