State Estimation by Using Particle Filter based on Quantized Outputs

Abstract

This paper considers the state estimation problem for nonlinear systems based on the quantized outputs. This problem plays an important role in achieving higher control performance when we use low-resolution sensors or networked control systems. First, the problem is formulated in a general setting, which could deal with a broad class of nonlinear systems in the presence of non-Gaussian noises. Second, it is proposed to apply the particle filter, which does not depend on linearity of the target systems nor Gauss noises, for the state estimation subject to quantized outputs. Numerical examples are given to demonstrate its effectiveness, where it is also shown how to deal with a class of uncertainty of the target systems.