PID-Based Position and Trajectory Control of a Four-Wheeled Omnidirectional Robot Using Robot Operating System (ROS)

Abstract

Precise position control in omnidirectional mobile robots is essential for applications in industrial automation and competitive robotics, including the Indonesian Robot Contest (Soccer Wheeled Middle Size League Division). This study aims to develop and evaluate a position control system for a four-wheeled omnidirectional robot using a PID controller through both simulation and physical implementation within the Robot Operating System (ROS) framework. The research was conducted by designing a robot model with four omniwheels arranged at 90° angles, integrating sensor fusion using an MPU6050 gyroscope, rotary encoders, and magnetic encoders to provide real-time position feedback (x, y, θ). PID parameters were tuned using Ziegler-Nichols and trial-and-error methods and tested across five trajectory scenarios: straight line, L-pattern, square, triangle, and maneuver paths. Simulation results using ROS-Gazebo demonstrated optimal performance with 1.60% overshoot, 0.732 s rise time, 2.380 s settling time, and 0.0018 m/s steady-state error. Physical implementation revealed that trial-and-error tuning provided the most balanced performance with 0.684 s rise time, 3.29% overshoot, and 2.872 s settling time, showing better adaptability to real-world disturbances compared to the more aggressive Ziegler-Nichols response. The PID controller effectively reduced overshoot from 8.85% to 3.29% and RMSE from 0.7025 to 0.4279 m/s compared to uncontrolled operation. These findings demonstrate the effectiveness of the proposed control system in achieving accurate positioning and trajectory tracking, with strong consistency between simulation results and real-world testing results. This research contributes to quality education in robotics (SDG 4), supports innovation in industrial automation (SDG 9), and establishes a foundation for collaborative research and development in robotic systems (SDG 17).