Fault-tolerant filter design with quantized measurements

Abstract

In this paper, a novel approach is proposed for designing fault-tolerant filters for linear discrete-time systems whose measurements undergo quantization. Two types of faults are considered: internal faults which are modelled by uncertainties in state space parameters; and actuator faults such as offset and stuck which are modelled as an unknown disturbance on the states. The proposed filter consists of two robust filters which operate in two different zones. These zones are defined based on the effects of the quantization on the residual signal between the quantized measurement and the one generated by the filter. Using an LMI-based approach with prescribed H∞ performance, stability LMIs are derived for the proposed dual-zone filter and finally the results are validated through simulation examples.