Performance Comparison between Position Controllers for a Robotic Arm Manipulator

摘要

This paper compares five PID controller architectures for robotic manipulator position control, addressing the challenge of maintaining performance under varying inertial loads while providing accessible implementations for research and education. The five PID controller architectures for a three degrees-of-freedom SCARA manipulator position control are a basic Proportional-Derivative (PD), PD with Feed-Forward (FF), Parallel PD-PI-FF, Cascade PD-PI-FF, and Cascade PD-PI - FF with dead zone (DZ) compensation. The controllers were evaluated under varying inertial loads to assess robustness, extending beyond previous work's idealized conditions. Results show advanced configurations reduced errors by up to 64% compared to the baseline PD, with Parallel-FF achieving optimal dynamic performance and Cascade-FF-DZ excelling in steady-state control. The Feed-Forward addition enhanced tracking performance, while DZ compensation effectively eliminated limit cycles. The work provides open-source implementations and simulation environments, supporting research reproducibility and educational applications in robotics control.