Adaptive reliable H∞ filtering against sensor failures

This paper is concerned with the adaptive reliable H-infinity filtering problem against sensor failures for continuous-time linear systems. An adaptive H-infinity performance index is defined to describe the disturbance attenuation performance of systems with time-varying parameter estimations. By combining the linear matrix inequality (LMI) approach for H-infinity filter design and adaptive method, a method of designing adaptive reliable H-infinity filters is proposed, where the filter parameter matrices are updated automatically to compensate the sensor failures effects on systems based on the online estimations of eventual faults. It is shown that the design condition for the newly proposed adaptive H-infinity filtering is more relaxed than that for the, traditional fixed gain filter design based on LMI approach. An illustrative example and its simulation results are given to illustrate the effectiveness of the proposed design method.