Autonomous fault management has recently been identified as a key technology for future space missions in NASA's 2010 technology roadmap. An innovative approach to vehicle health monitoring (VHM) is proposed using constraint suspension with parity space and hypothesis testing techniques. The aim is to explicitly model and use information about sensor and process noise to create a VHM algorithm that is designed for tunable, predictable reliability. The proposed algorithm is described and preliminary simulation and analytic results are reported that demonstrate an improvement over constraint suspension using fixed threshold node comparison.
展开▼
