Linear Quadratic Regulator Path Tracking for Omnidirectional Robots in High-Dynamic Environments

• João Victor Lourenço Aguiar Centro Universitário FEI
• Leonardo Da Silva Costa Centro Universitário FEI
• Flavio Tonidandel Centro Universitário FEI

Abstract

In the high-dynamic Small Size League (SSL) environment of RoboCup, there is a high need for precise path tracking to ensure well-played matches. Thus, this work proposes using the Linear Quadratic Regulator (LQR) to achieve better tracking control for the omnidirectional SSL robots in this dynamic environment. A comparison between the LQR and the PID controller, a usual controller in the league, will verify the proposed approach. A square-based path and a spline path, with two different speeds, are used to test the controllers. The metrics used to evaluate the tests are overall navigation time and mean lateral, orientation, and velocity error. The results showed that the LQR is a good solution for path tracking control for omnidirectional robots in high-dynamic environments. In lower speeds, the LQR was better than PID analyzing the mean lateral, orientation, and speed error. In addition, although LQR had better results than the PID controller analyzing total time, orientation, and speed error, both controllers still were not able to track the paths with a low error in higher velocities.