Collaborative Operation of Artificial Intelligence and Mechanical Arm Based on Transformer-XL

Authors

  • Wenzhi Zhang Department of Mechanical and Electronic Engineering, Sichuan Vocational and Technical College of Communications, Chengdu 611130, Sichuan, China
  • Min Xia College of Engineering and Technology, Nanchang Vocational University, Nanchang 330000, Jiangxi, China
  • Biao Yang School of Mechanical and Electrical Engineering, Gongqing Vocational College of Science and Technology, Jiujiang 332000, Jiangxi, China

Abstract

In response to the problems of insufficient flexibility, slow response speed, and poor long-term dependency processing ability in traditional mechanical arm collaborative operations, this article combined Transformer-XL and deep Q-learning (DQL) to improve the adaptability and decision-making efficiency of mechanical arms in complex dynamic environments, thereby achieving more efficient and accurate collaborative operations. Firstly, the collected sensor data can be cleaned and normalized to improve the effectiveness of model training. Secondly, building deep learning models based on the Transformer-XL architecture can improve the ability to capture long-term dependency relationships. Then, based on the DQL algorithm, the decision-making process of the mechanical arm in dynamic environments can be optimized. Finally, the output of Transformer-XL can be combined with DQL to form an efficient collaborative control strategy. The experimental results show that compared with traditional control methods, DQL (combined with Transformer-XL) exhibits significant advantages in various task environments. In more complex dynamic obstacle avoidance tasks, the decision time is only 220 milliseconds, while maintaining a success rate of 88%. It demonstrates faster response speed and higher success rate in complex dynamic environments. At the same time, it still maintains a success rate of 60% with a delay of 90 seconds, demonstrating stronger adaptability in dynamic environments compared to traditional control strategies. The outstanding performance in high dynamic environments also validates the reliability of the research content.

Keywords: Mechanical Arm; Transformer-XL Architecture; Deep Q-learning; Normalization Processing; Control Strategy

Cite As

W. Zhang, M. Xia, B. Yang, "Collaborative Operation of Artificial Intelligence and Mechanical Arm Based on Transformer-XL",
Engineering Intelligent Systems, vol. 33 no. 5, pp. 499-511, 2025.

Downloads

Published

2025-09-01

Issue

Vol. 33 No. 5 (2025): International Journal of Engineering Intelligent Systems

Section

Article

License

The submission of a paper implies that, if accepted for publication, it will not be published elsewhere in the same form, in any language, without the prior consent of the publisher. Before publication, authors are requested to assign copyright to CRL Publishing Ltd. This allows CRL to sanction photocopying, and to authorize the reprinting of issues or volumes according to demand. Authors' traditional rights will not be jeopardized by assigning Copyright in this way, as they retain the right to reuse the material following publication, and to veto third-party publication.