Fault detection and isolation for stochastic linear hybrid systems

A model-based fault detection and isolation (FDI) scheme is proposed for a stochastic linear hybrid system (SLHS) which has both discrete state dynamics and linear continuous state dynamics. We first propose a residual generation filter for the SLHS. It is shown that this filter generates a residual vector with zero mean and a known covariance when the SLHS matches the plant dynamics. Based on the known statistical properties of the residual, we then design a FDI scheme based on multiple hypothesis testing. We illustrate an application of the FDI scheme to an aircraft autolanding system.