A Testbed for Real-Time Autonomous Vehicle PHM and Contingency Management Applications

Abstract

Autonomous unmanned vehicles are playing an increasingly important role in support of a wide variety of present and future critical missions. Due to the absence of timely pilot interaction and potential catastrophic consequence of unattended faults and failures, a real-time, onboard health and contingency management system is desired. This system would be capable of detecting and isolating faults, predicting fault progression and automatically reconfiguring the system to accommodate faults. This paper presents a robotic testbed that was developed for the purpose of developing and evaluating real-time PHM and Automated Contingency Management (ACM) techniques on autonomous vehicles. The testbed hardware is based on a Pioneer 3-AT robotic platform from Mobile Robots, Inc. and has been modified and enhanced to facilitate the simulations of select fault modes and mission-level applications. A hierarchical PHM-enabled ACM system is being developed and evaluated on the testbed to demonstrate the feasibility and benefit of using PHM information in vehicle control and mission reconfiguration. Several key software modules including a HyDE-based diagnosis reasoner, particle filtering-based prognosis server and a prognostics-enhanced mission planner are presented in this paper with illustrative experimental results. This testbed has been developed in hope of accelerating related technology development and raising the Technology readiness level(TRL)ofemergingACM techniques for autonomous vehicles.