Positioning Error Compensation for Fully-Gear-Driven Robotic Manipulator Based on Visual Calculation

Zhi Liu, Jituo Li, Wentang Chen, Tingxi Xue, Pei Yuan, Dingcan Jin, Jiefeng Jin, Yifeng Zhang

Abstract: In response to the problem of excessive positioning errors in Fully-Gear-Driven Robotic Manipulator(FGDRM) that hinders automated operations, this paper proposes a positioning error compensation technique for FGDRM based on visual calculation. A visual measurement system is established, in which methods of multi-cameras arrangement optimization and FGDRM pose calculation are proposed. Additionally, an online positioning error compensation approach based on a fuzzy neural network PID algorithm is proposed. Experimental results demonstrate that the positioning accuracy for point-to-point, straight-line, and arc trajectories is improved by over 93%, with the mean positioning error reduced from 73.39 mm to 2.93 mm. This technique provides a support for the automation of FGDRM operations.