Spacecraft Navigation Using a Robust Multi-Sensor Fault Detection Scheme

Redundant sensor networks of inertial measurement units (IMUs) provide inherent robustness and redundancy to a navigation solution obtained by dead reckoning the fused accelerations and angular velocities sensed by the IMU. However, IMUs have been known to experience faults risking catastrophic mission failure creating large financial setbacks and an increased risk of human safety. Robust on-board fault detection schemes are developed and analyzed for a multi-sensor distributed network specifically for IMUs. Simulations of a spacecraft are used to baseline several cases of sensor failure in a distributed network undergoing fusion to produce an accurate navigation solution. The presented results exhibit a robust fault identification scheme that successfully removes a failing sensor from the fusion process while maintaining accurate navigation solutions. In the event of a temporary sensor failure, the fault detection algorithm recognizes the sensors\u27 return to nominal operating conditions and processes its sensor data accordingly