Reconfigurable control of mobile robot based on a robust estimator

This paper introduces a procedure for the design of reconfigurable linear quadratic (LQ) state-feedback control tolerant to actuator fault. In fact, this work is based on a fast adaptive fault diagnosis observer. We consider the Unknown Input Observer (UIO) which is subsequently used for a robust fault detection scheme and also as an adaptive detection scheme for an additive actuator fault. Stability of the adaptive estimation is provided by a Lyapunov function ending with solving the Linear Matrix Inequalities (LMI). Therefore, thanks to the loop correction based on this novel estimation design, the process performances will be improved. As an example of simulation, a linear model describing an unicycle robot is proposed to illustrate the theoretical results.