Abstract
In this study, we propose a collision avoidance control method using the model predictive control for omnidirectional mobile robots which have convex polygonal shapes. It is possible to modify the conventional oval potential field method to consider the shape of the robot, but this potential field may have local minimum points. Thus, we execute the path planning based on the conventional oval potential field method within the predictive interval to reduce local minimum points. Then, the model predictive control which has an artificial potential function term considering the shape of the robot is executed. We use this planned path as the reference; the inputs of the model predictive control reduces local minimum points which comes from the artificial potential field considering the shape of the robot. Hence, the proposed method achieves both consideration of the shape of the robot and reduction of local minimum points. Moreover, we conduct numerical simulations and experiments to verify that the proposed method can prevent local stationary points and achieves the collision avoidance, while it takes into consideration the shape of the robot.
