CAREER: Fault Diagnostics for Automated Vehicle Applications

拉杰什·拉贾马尼，明尼苏达大学#9983946号提案自动化车辆应用的故障诊断这个项目专注于开发故障诊断系统，该系统可以监控自动车辆上所有传感器和执行器的健康状况，并确定发生的任何故障的来源。对半自动车辆(如在自适应巡航控制和车道保持类型应用中的应用)以及专用自动公路系统上的全自动车辆进行了研究。首先，由于车辆动力学模型可能是高度非线性的，基于非线性观测器设计结果，提出了一种针对非线性系统的系统故障诊断方法。其次，解决了为测量车辆间变量和车-路变量的传感器获得分析冗余的问题。这项任务的困难在于，这些传感器测量中的每一个都涉及第二个动态系统，而另一个动态系统的传感器无法访问。该项目的试验性工作计划包括在MnRoad研究设施对一辆沃尔沃卡车和一辆Navistar牵引车进行整个故障诊断系统的试验实施。该项目涉及车辆/公路自动化的一个重要方面。为了使车辆自动化成为可能，开发可靠的故障诊断和故障处理系统是必要的。该项目的教育目标包括教科书和课程开发。

Rajesh Rajamani, University of Minnesota Proposal # 9983946 Fault Diagnostics for Automated Vehicle Applications ABSTRACTThis project focuses on the development of fault diagnostic systems that can monitor the health of all the sensors and actuators on an automated vehicle and identify the source of any fault that occurs. Both semi-automated vehicles (such as in adaptive cruise control and lane-keeping-type applications) as well as fully automated vehicles on dedicated automated highway systems are investigated.Two key research challenges are addressed in the project. First, since vehicle dynamic models can be highly nonlinear, a systematic fault diagnostics methodology is developed for nonlinear systems based on nonlinear observer design results. Second, the issue of obtaining analytical redundancy for sensors that measure inter-vehicle and vehicle-road variables is addressed. The difficulty in this task arises from the fact that each of these sensor measurements involve a second dynamic system, with no access being available to sensors of the other dynamic system. The experimental work plan in the project includes experimental implementation of the entire fault diagnostic system on a Volvo truck and a Navistar tractor-trailer at the MnRoad research facility. The project addresses an important aspect of vehicle/highway automation. The development of reliable fault diagnostic and fault handling systems is necessary in order to make vehicle automation viable. Educational objectives of the project include textbook and curriculum development.