Discrete-time LPV H2 observer for vehicle model-based state observer

초록

In this paper, we propose the novel scheme to design a vehicle model-based state observer. The observer problem for vehicle state is reformulated into the linear parameter varying (LPV) systems which depend on online accessible time-varying parameters. In vehicle state estimation problem using global positioning system, we consider the measured course angle as varying parameter. The observer gain scheduling is determined using H2 filter based on linear matrix inequality (LMI) approach. The state observer gain using a discrete-time H2 filter could minimize an error from a disturbance to a state. The proposed method was demonstrated via performance comparison with EKF which is most popular method through experimental test.