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

电子化是汽车自动化和智能化的基础，需要部署越来越多的车载传感、通信和计算服务。然而，车辆网络化也带来了新的问题，包括潜在的软件漏洞、安全漏洞和错误的传感器读数，这些都可能降低车辆的可靠性并扰乱汽车行业。为了获得最佳性能，车辆诊断必须考虑跨物理和网络领域的交互作用、对易受攻击的车载网络的依赖以及无法处理未知异常。该奖项将支持基础研究，以解决使用汽车电池的车辆的网络异常问题。该方法将利用不同学科的主题，包括车辆系统、电池管理、数据分析和图论。这一结果将增强当前车辆的诊断能力，从而使汽车生态系统中的各方受益，从汽车制造商到车主。因此，这项研究的结果将使美国经济和社会受益。此外，这种多学科的研究将有助于扩大学生对工程和计算机的参与，特别是来自代表性不足的群体的参与。电池驱动的诊断系统可以克服现有解决方案的所有上述局限性，具有值得信赖、普遍适用于所有车辆、在整个车辆寿命内可靠的优势。然而，要充分发挥其应用潜力，还需要克服一些科学障碍。为此，本研究将使用网络物理方法对车辆运行和电池功率之间的依赖关系进行建模，使用两层图模型基于这些依赖关系对车辆进行抽象，并使用该图以四个渐进步骤指导车辆异常的诊断：a)以电池作为信任根检测异常，b)验证检测到的异常以减少虚警，c)通过图分解识别故障车辆模块，以及d)通过信息恢复缓解异常，以减少其对车辆运行的负面影响。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

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.