Robust fault detection of mobile robots using an ellipsoid bounding identification algorithm

This paper presents a set-membership based robust fault detection approach for mobile robots with dynamic linear in some fault-related parameters. In the framework of set membership identification, the model uncertainties are considered and modeled as unknown-but-bounded noises. A new set-membership identification algorithm using bounding ellipsoid for linear MIMO system derived from robot's dynamics is proposed to obtain the feasible set of parameters, which is supposed to be compatible with the model uncertainties assumptions. The fault detection scheme is established by consistency check of the feasible set in a robust way to distinguish real fault with model uncertainties. Simulation results are given to demonstrate the effectiveness and robustness of the proposed method.