Title: Dexterity control of multi-arm sorting robot based on machine learning

Authors: Linyan Pan

Addresses: Intelligent Manufacturing College, Anhui Wenda University of Information Engineering, Hefei, Anhui 230000, China

Abstract: In order to overcome the problems of large dexterity control error of manipulator joint and poor sorting and positioning accuracy, this paper designs a dexterity control method of multi manipulator sorting robot based on machine learning. Firstly, the attitude of the multi manipulator coordinate system on the rigid body is obtained. Secondly, the translation matrix is constructed by using the translation transformation method. Then, the rotation matrix is constructed to determine the inverse motion law of the robot. Finally, determine the dexterity parameters of the manipulator joint, introduce the machine learning algorithm to calculate the dexterity parameter control error, and correct the error through the activation function to complete the dexterity control. The experimental results show that the error of this method is always less than 0.1% and the positioning accuracy is higher than 90%, which shows that the dexterity control effect of this method is good.

Keywords: machine learning: multi-manipulator; robot; dexterity; translation transformation; rotation matrix; activation function.

DOI: 10.1504/IJMTM.2024.137387

International Journal of Manufacturing Technology and Management, 2024 Vol.38 No.1, pp.81 - 94

Received: 26 May 2022
Accepted: 01 Sep 2022
Published online: 15 Mar 2024 *