Sliding Mode Integral Observers for Sensor Faults Detection and Isolation in Nonlinear Systems

Abstract

In this paper, a new approach to sensor faults detection and isolation for a class of nonlinear systems is proposed. Through introducing a new state variable, an augmented system can be constructed to treat sensor faults as actuator faults in the form to avoid amplification of sensor faults. Then a bank of sliding mode integral observers (SMIO) are utilized to isolate all possible sensor faults and the so-call equivalent output error injection signals estimate them. Based on the Lyapunov stability theory, a sufficient condition for sensor faults isolation is derived in the form of linear matrix inequality (LMI). Finally, a simulation example is presented to show that the unknown sensor faults are detected and isolated accurately, and the equivalent output error injection signals can estimate the sensor fault signal with high precision.