Control of a robot-trailer system using a single non-collocated sensor

Abstract

In this paper, the authors develop an observer and controller for a robot and trailer system in which only a single non-collocated position measurement is used. A linearized state-space model of the system is derived using kinematic equations that have previously proven sufficient for state feedback control. Optimal observer gains (LQG) and controller gains (LQR) are calculated using known process and measurement noise variances. Simulations and experiments both confirm that the LQG/LQR system controls the nonlinear system in path tracking. Control based on state estimation yields performance that is comparable to a more heavily-instrumented full-state feedback system previously used on the robot-trailer system.