A Finite State Approximation Based Approach for Motion Planning of Discrete-Time Linear System

摘要

We consider the problem of motion planning of discrete-time linear systems. Our mission is to design a controller, which drives the state to the destination without entering a prescribed unsafe region. The approach is based on a finite state approximate system, whose state space is a finite set. In order to design the controller, the ρ/μ approximation, which was proposed by Tarraf [IEEE Trans. AC, 56-12 (2012)], is applied to design the controller. The merit of ρ/μ approximation is that we can design a controller by finding a control policy for the approximate system. A possible way to search the control policy for the approximate system is presented, which is based on the modification of the method for the benchmark problem.