Model-Based Design for Vehicle Health Monitoring

NASA’s 2012 Technology Roadmap identifies autonomous vehicle health monitoring (VHM) as a key technology for future space missions. This paper describes the demonstration of an algorithm for constraint-based VHM software. The VHM approach, referred to as Stochastic Constraint Suspension (SCS), uses system models, constraint suspension, and hypothesis testing to detect and isolate faults, given dynamic estimates of sensor uncertainty and process noise. Through stochastic analysis on a simulated aerospace vehicle subsystem, we design, tune, and verify the success rate of the health monitoring algorithm against identification and isolation metrics.