Model Predictive Control for Omnidirectional Small Size Robot with Motor and Non-Slipping Constraints

Abstract

The Robocup Small Size Soccer League is a widely known Robotics league around the world. One important competition done in Robocup is the Small Size Soccer League (SSL). The path following problem is important to score goals in this competition. It is also necessary a speed control algorithm embedded in the low level. The teams mostly use embedded linear speed controllers. However, the robot has constraints not considered by linear control theory. The main goal of this paper is to derive a mathematical description of the robot's model and propose a Model Predictive Controller that considers these constraints, getting near to an optimal approach for the problem. The speed controller achieved shows a system that uses the maximum acceleration available throughout the transient. Besides this, the controller obtained slows down the system to avoid slippages. The slippages could drastically reduce the robot's performance when following paths. Finally, the linear physical model obtained for the constraints opens up opportunities for implementing faster algorithms using e.g. Multi Parametric Programming.