Filtering method for nonlinear systems with constraints

Abstract

A constrained filtering method is proposed to deal with the filtering problems for nonlinear systems with constraints. The problem is converted to a sequence of recursive estimation problems in which the system equations and constraint conditions are treated as pseudo-measurements. To resolve the singularity problem arising from the constraints, a modified maximum-likelihood method for nonlinear systems is developed. The simulation results from the application of the proposed scheme to the target tracking problem shows that the constrained filtering method can enhance the performance of filter design significantly.