Calibration of a six-axis parallel manipulator based on BP neural network

Purpose: This paper aims to provide a simple and flexible calibration method of parallel manipulators for improving the position accuracy only using partial pose information. Design/methodology/approach: The overall idea of this method is to use BP neural network to fit the relationship between calibration parameters and measurement parameters and then adjust calibration parameters according to measurements. Findings: The calibration method significantly improves the position accuracy of the six-axis parallel manipulator. Simulation shows that the accuracy can be improved by increasing the number of positions consisted of samples to train BP neural network, and when the position number is increased, the descent velocity of fitting error is decreased. Originality/value: The method is general for various parallel mechanisms and simple for measurement process. It can be applied to the calibration of various mechanisms without analyzing the mathematical relationship between measurements and calibration parameters. The measurement parameters can be flexibly selected to simplify measurement process, which saves calibration cost and time.