No Cover Image

Journal article 634 views 132 downloads

Recurrent Neural Networks-Based Collision-Free Motion Planning for Dual Manipulators Under Multiple Constraints

Jinglun Liang, Zhihao Xu, Xuefeng Zhou, Shuai Li Orcid Logo, Guoliang Ye

IEEE Access, Volume: 8, Pages: 54225 - 54236

Swansea University Author: Shuai Li Orcid Logo

  • 53940.pdf

    PDF | Version of Record

    This work is licensed under a Creative Commons Attribution 4.0 License.

    Download (2.45MB)

Check full text

DOI (Published version): 10.1109/access.2020.2981688

Abstract

Dual robotic manipulators are robotic systems that are developed to imitate human arms, which shows great potential in performing complex tasks. Collision-free motion planning in real time is still a challenging problem for controlling a dual robotic manipulator because of the overlap workspace. In...

Full description

Published in: IEEE Access
ISSN: 2169-3536
Published: Institute of Electrical and Electronics Engineers (IEEE) 2020
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa53940
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract: Dual robotic manipulators are robotic systems that are developed to imitate human arms, which shows great potential in performing complex tasks. Collision-free motion planning in real time is still a challenging problem for controlling a dual robotic manipulator because of the overlap workspace. In this paper, a novel planning strategy under physical constraints of dual manipulators using dynamic neural networks is proposed, which can satisfy the collision avoidance and trajectory tracking. Particularly, the problem of collision avoidance is first formulated into a set of inequality formulas, whereas the robotic trajectory is then transformed into an equality constraint by introducing negative feedback in outer loop. The planning problem subsequently becomes a Quadratic Programming (QP) problem by considering the redundancy, the boundaries of joint angles and velocities of the system. The QP is solved using a convergent provable recurrent neural network that without calculating the pseudo-inversion of the Jacobian. Consequently, numerical experiments on 8-DoF modular robot and 14-DoF Baxter robot are conducted to show the superiority of the proposed strategy.
Keywords: Motion planning, dual robotic manipulators, dynamic neural networks, zeroing neural networks, redundant resolution
College: Faculty of Science and Engineering
Start Page: 54225
End Page: 54236

Similar Items