A Novel Small Fault Detection Scheme Based on Local Approach and Divided Difference Filter for Inertial Sensors in Unmanned Aerial Vehicles

Abstract

This article deals with the problem of fault detection for inertial sensors of unmanned aerial vehicles (UAVs). Novel fault detection (FD) method for nonlinear systems is proposed using the residuals generated by the second-order divided difference filter (DD2 filter) and the local approach. DD2 filter is based on the polynomial approximation of the nonlinear transformations obtained with a particular multi-dimensional extension of Stirling's interpolation formula. As the local approach is a powerful statistical tool for detecting changes in the mean of a Gaussian process, it is used to detect faults from residuals obtained from the DD2 filter. The key feature of the proposed method is that it can successfully and quickly detect small faults under noisy conditions. The comparison between local approach and generalized likelihood test approach is introduced to illustrate the effectiveness of the proposed method. To demonstrate the implementation and performance of the proposed technique, it is applied to detect inertial sensor faults in UAVs system.