Search

In path following control, the forward velocity can be seen as an additional degree of freedom. However, many path following control laws assume that the forward velocity is a constant and actuators are ideal, i.e., no actuator constraints. Unfortunately, these controllers cannot be realized in real-world implementation. Thus, we propose an extension to an existing path following controller in order that the forward velocity can be varied according to the robot’s posture errors and the desired forward velocity. This can be done by adopting a receding horizon scheme, where the optimization problem is solved online along some look-ahead time intervals. Furthermore, we treat boundaries of wheel speed and acceleration as constraints in the optimization problem. As a result, a mobile robot can slow down when making a sharp turn and it can speed up when being behind a reference point. Simulation tests have been conducted in order to illustrate the effectiveness of our proposed scheme.