Development of an All-condition Dependable Vehicle Health Monitoring System

汽车工业正处于汽车技术进步的转型期。在低成本传感器、高处理能力和无处不在的实时信息访问方面取得的重大进展正在使车辆更安全、更智能，从而提供更好的控制，为全自动驾驶铺平道路。从以人为中心的决策到以机器为中心的决策，这种前所未有的演变增加了对算法控制车辆的依赖，导致了与可靠性、准确性、健康和安全相关的重大担忧。今天的高级驾驶辅助系统（ADAS）通过减少人为错误来提高安全性，但仍然需要驾驶员的注意力和互动来转移系统控制，并在不明确的情况下做出决定。严格的诊断算法对车辆主动安全系统的安全和稳健性能至关重要，更重要的是对全自动驾驶系统至关重要。在当前的系统中，重点是传感器和执行器的故障检测，而不是状态估计或控制系统。拟议的研究项目将开发对自动驾驶和ADAS至关重要的车辆健康监测系统，包括安全关键估计和控制单元以及动力总成系统的准确性和性能监测。估计和控制模块的健康监测工作非常有限，尽管它们在配备ADAS和自动驾驶的车辆中至关重要，因为驾驶员的注意力不需要完全集中。电力传动系统的任何部分或全部故障都可能严重危及车辆的安全，因此需要对其进行强大的实时监控。这项研究将为所有条件下实用可靠的自动驾驶技术开发必要的知识和技术，使自动驾驶和整体道路旅行更安全。这种跨学科的研究将提供一个充满活力的培训环境，让学生和工程师接触到广泛的学科和技术，并有机会在不断发展的加拿大运输行业中发展实际应用的研究技能。

The automotive industry is in a transformational period of vehicle technology advancement. Significant progress in low-cost sensors, high processing power, and ubiquitous real-time information access are making vehicles safer and more intelligent, in turn providing better control and paving the way to fully automated driving. This unprecedented evolution in transferring human-centered decisions to machine-centered decisions increases reliance on algorithms to control vehicles, leading to significant concerns related to reliability, accuracy, health and safety. Today's Advanced Driver Assistance Systems (ADAS) increase safety by reducing human error, but still require driver attention and interaction to transfer the system control and make decisions in cases of ambiguity. Rigorous diagnosis algorithms are essential for safe and robust performance of vehicle active safety systems and, more importantly, for fully automated driving systems. In current systems, the focus is on sensor and actuator fault detection and not on state estimation or control systems. The proposed research project will develop vehicle health monitoring systems that are critical to automated driving and ADAS which include accuracy and performance monitoring of safety critical estimation and control units, and powertrain systems. There are very limited works in health monitoring of estimation and control modules although they are critical in vehicles equipped with ADAS and automated driving in which driver's attention is not fully required. Any full or partial failure in electric powertrain could seriously endanger the safety of the vehicle for which a robust real-time monitoring is needed. This research will develop the necessary knowledge and technologies for practical and dependable automated driving technology in all conditions, making automated driving and overall road travel safer. This interdisciplinary research will provide a dynamic training environment, exposing students and engineers to a broad range of disciplines and technologies, as well as the opportunity to develop research skills with practical applications in the evolving Canadian transportation industry.