In this report, a preliminary design of a health monitoring system for automated vehicles is developed and results of tests in a high-fidelity nonlinear simulation are reported. Two car models are described: low- dimensional linear models for fault detection filter design and high fidelity nonlinear model is used for evaluation and to obtain the linear models used for design. The faults and the fault detection filters and system are then described. An evaluation of the performance of the fault detection filters in a nonlinear simulation are presented. A fault detection filter residual processing system, using a neural network, is then described.

This report continues work on the design of a health monitoring system for automated vehicles. The approach is designed to fuse data from dissimilar instruments using modeled dynamic relationships and fault detection and identification filters. Issues relating to sensor models, output separability, steady-state fault persistence and the spectral content of sensor faults are considered.

This work focuses on vehicle fault detection and identification. It also describes a vehicle health management approach based on analytic redundancy. It describes a point design of fault detection filters and parity equations that are developed for the vehicle longitudinal mode. A preliminary design of a range sensor fault monitoring system is outlined as an application of a new decentralized fault detection filter. This system combines dynamic state information already generated by the existing filter designs with inter-vehicle analytic redundancy