Safety-critical systems refer to systems whose failure or malfunction may result in death or serious injury to people, loss or severe damage to equipment, or environmental harm. One key issue is whether it is safe to operate wireless devices for monitoring, control, and coordination in these systems, or whether current wireless devices can safely coexist with these critical devices? This project develops novel and holistic solutions to monitoring, policing and assessing the coexistence of wireless devices in and near safety-critical systems. Intellectually, the project advances the state-of-the-art in both algorithm design as well as experimental systems related to safety-critical applications through 1) the design of a distributed passive monitoring system that captures cross-layer information regarding radio frequency activities, 2) the development of device profiling and identification algorithms for characterizing and tracking active devices, 3) the development of wireless advisory to identify and make recommendations on the proper responses to potentially safety hazardous conditions, and 4) validation through real-world experiments.
Addressing the co-existence issues of wireless devices for safety-critical applications has far-reaching societal impacts. Ensuring the safe operation of wireless technologies in or near safety-critical applications can automate the process, reduce human errors, and avoid safety hazards. The interdisciplinary nature of this transformative research contributes to the development of the nation's future workforce by ensuring that undergraduate, pre-doctoral, and postdoctoral students receive the didactic and research experiences necessary to engage in integrative and team approaches to solve complex problems.
Safety-critical systems refer to systems whose failure or malfunction may result in death or serious injury to people, loss or severe damage to equipment, or environmental harm. One key issue is whether it is safe to operate wireless devices for monitoring, control, and coordination in these systems, or whether current wireless devices can safely coexist with these critical devices? This award develops novel and holistic solutions to monitoring, policing and assessing the coexistence of wireless devices in and near safety-critical systems. The goal of the project is to advance the state-of-the-art in both algorithm design as well as experimental systems related to safety-critical applications through 1) the design of a distributed passive monitoring system that captures cross-layer information regarding radio frequency activities, 2) the development of device profiling and identification algorithms for characterizing and tracking active devices, 3) the development of wireless advisory to identify and make recommendations on the proper responses to potentially safety hazardous conditions, and 4) validation through real-world experiments. During the life of the award, we were engaged in several interdisciplinary and transformative activities toward accomplishing the goals. Developing new distributed sequential learning algorithms for wireless monitoring. We have identified the scalability problem of previous learning algorithms and proposed several approximate algorithms that are amenable to distributed implementation. Conducting spectrum measurement in hospital environments. We have collected spectrum data in different settings in a local hospital including waiting area, doctor's office and outside intensive care unit. Developing an indoor tracking solution for mobile users. We devise a nonparametric Bayesian clustering approach to identify recurring places a mobile user visits indoor. Extending to D2D communications using the overlay communications so as to improve the coverage and reliability. The specific objectives are to develop efficient solutions to monitoring the spectrum usage, existing wireless devices and users. From the algorithm point of view, we need to construct schemes that are more robust to the missing data and errors. Moreover, the algorithms should have no prior information and can learn by themselves. From the prototyping point of view, we need to have get traces in the real environment, and implement the algorithms in USRP2 and smart mobile phones. We have found the following significant results: The approximate learning algorithm for wireless monitoring is significant lower computation complexity and is suitable for distributed implementation. We also devise sequential learning algorithm for wireless monitoring with fast convergence. We are able to identify major sources of wireless transmissions in the local hospital environment. We collect more comprehensive spectrum data and conduct survey on wireless device usage in clinical settings. The indoor tracking solution can provide zone-level identification of mobile users using minimum number of samples while maintaining high accuracy. Improve co-existence among heterogeneous wireless devices. For broader impact, the project has involved 2 graduated Ph.D., 1 Ph.D. candidate and 2 MS graduated students. The training provided through the project has trained the students in both theoretical foundations as well as experimental tools, and contributed to prepare the students for their future career. The results have been disseminated through 7 conference presentations (including one best paper award in IEEE Wireless Communications and Networking Conference, 2013), 8 journal publications, research talks and tutorials. Electronic materials have been made available through our research website. We also incorporate the research in the existing curriculum at University of Houston. Finally, outreach activities will be directed to high school students and increase the participation of women and minority in science and engineering.
