Two Heuristic PID Tuning for a 4-DOF Robot Arm Control

Abstract

This paper discusses the development of closed-loop position control for the joint motor of a 4-DOF robot arm. In developing a practical robot, types of input and control algorithm are important elements in the control system. The objectives of this paper are to select an input type that could produce smooth arm’s motion and to apply PID control for the joints. Step and parabolic inputs have been tested as reference trajectory where the former used step signal as the constant value of motor angle while the latter requires calculation using spline interpolation method based on initial and end points of each joint motion given by user. From the latter, a parabolic curve input which consists of point-to-point values calculated for each 1ms sampling time was produced. Meanwhile, the PID position control tuned by trial and error and Ziegler Nichols (Z-N) methods were carried out and measured for comparison in terms of steady state error, overshoot and response time in real-time experiments. It could be observed that step input has caused abrupt movements and vibration to the arm body compared to the parabolic input that moved the arm smoothly to imitate motion like human. Meanwhile, the intuitive values of trial and error method were found to produce not much different with the Z-N tuned gains but comparatively produced small error against the parabolic input reference during the transient state. The result from this study is useful in the future to ensure position accuracy of robot joints when attending object for manipulation.