Fault detection and isolation enhancement of an aircraft attitude and heading reference system based on MEMS inertial sensors

Abstract

The process of enriching a multisensor avionic strap-down attitude estimation unit with fault detection and isolation (FDI) capabilities is presented. Compact solid-state sensors are integrated in an autonomous emergency guidance system for aircrafts. Optimal attitude estimate is provided by a 9-state extended Kalman filter for fusion of complementary inertial data. Observerbased residual generation can be consequently added with negligible computational overhead. On-line FDI of the different sensors is analyzed and achieved with minimal hardware and software modifications. The proposed detection scheme combines different approaches: physical redundancy (accelerometers), injection of spectrally-decoupled test signals (magnetic sensors) and analytical redundancy (gyros).