Diagnosing Vehicles Using Automotive Batteries as Physical Root-of-Trust

Cyberization is the foundation of automated and intelligent vehicles, requiring the deployment of ever-increasing onboard sensing, communication, and computing services. However, vehicle cyberization also creates new problems, including the potential for software bugs, security vulnerabilities, and erroneous sensor readings, which can degrade vehicle reliability and disrupt the automotive industry. For optimal performance, vehicle diagnostics must consider the interactions across the physical and cyber domains, the reliance on vulnerable in-vehicle networks, and the inability to address unknown anomalies. This award will support fundamental research to address the cyber anomalies of vehicles using automotive batteries. The approach will leverage topics from different disciplines, including vehicular systems, battery management, data analysis, and graph theory. The results will augment current vehicles’ diagnostic ability and thus benefit all parties in the automotive ecosystem, from automakers to car owners. As such, the results from this research will benefit the U.S. economy and society. In addition, this muti-disciplinary research will help broaden student participation in engineering and computing, especially from underrepresented groups. The battery-enabled diagnostic system can overcome all the above-stated limitations of existing solutions, with the advantages of being trustworthy, universally applicable to all vehicles, and reliable throughout the vehicle life. However, some scientific barriers are yet to be overcome to realize the full application potential. To this end, this research will model the dependency between vehicle operation and battery power using a cyber-physical approach, abstract the vehicle based on these dependencies using a 2-layer graph model, and use the graph to guide the diagnostics of vehicle anomalies with four progressive steps: a) detect anomalies using the battery as a root-of-trust, b) verify the detected anomalies to reduce false alarm, c) identify the faulty vehicle module via graph decomposition, and d) mitigate anomalies to reduce their negative impacts on vehicle operation via information recovery. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.