Self-Balancing Robot Control Optimization Using PSO

• Published in: 2020 5th International Conference on Control and Robotics Engineering (ICCRE) pp. 7 - 10
• Main Authors: Mendez-Flores, Efrain, Martinez-Galicia, E. Mauricio, Lozoya-Santos, Jorge de J., Ramirez-Mendoza, Ricardo, Morales-Menendez, Ruben, Macias-Hidalgo, Israel, Vargas-Martinez, Adriana, Molina-Gutierrez, Arturo
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.04.2020
• Subjects: Convergence; discrete systems; Mathematical model; Mobile robots; Optimization; PID; PSO; Robustness; Self-balancing robot; Tuning
• DOI: 10.1109/ICCRE49379.2020.9096470

Robotic systems are ideal testbeds for mechanical, electrical and controllers integration, in order to test disturbance resistance to some unmodeled parameters. This paper discusses the optimization of the PID controller gains through the Particle Swarm Optimization (PSO) algorithm, implemented in order to achieve a steady upright position of a two-wheeled robot. The description of the testbed, as well as the implementation of the controller with the optimized gains is detailed in the content of this work. Results in this work, show an improved behavior of the PID controller tuned through the PSO, compared to a classical tuning method (such as Ziegler-Nichols method). In spite that both controllers accomplished to maintain the pendulum in vertical position (0^{\circ}\, \pm 5^{\circ}), the optimized controller managed to dampen the oscillations around the upright position, which was traduced into more than 80% of improvement regarding convergence and disturbance resistance.