Study of fast orthogonal search for fault detection in multi-inertial sensors frame

A de-noise method using fast orthogonal search (FOS) is studied in this paper for improving multi-inertial sensors system fault detection. It is established that the sensor measurement noise obey Gauss white distribution, and a traditional low passed filter (LPF) is applied in a set of simulated inertial measurement unit (IMU) data to compare FOS performance. In contrast to LPF and unprocessed data, FOS exhibits a considerable performance although a heavier computation and model time are required.