Research on Obstacle Avoidance Motion Planning of Space Manipulator Based on Reinforcement Learning

Zixuan Zhang*, Wei Dong, Chunyan Wang, Jing Sun

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Making up for the lack of generalization and environmental simulation of traditional algorithms, a motion planning method of space manipulator based on reinforcement learning is designed. First, the standard Denavit-Hartenberg(DH) model of space manipulator is given. Further, combined with the characteristics of the space mission, the state space, action space and reward functions are designed. The Proximal Policy Optimization(PPO) is used as the framework to realize the motion planning task of the space manipulator. ISAAC GYM is chosed as the simulation platform to improve the training speed and strategy generalization ability through the setting of multiagent training and environment randomization at the same time. The simulation results show that the proposed method can realize the task of grasping the object by avoiding obstacles in the case of space microgravity, and the method has strong practicability and effectiveness.

Original languageEnglish
Title of host publicationAdvances in Guidance, Navigation and Control - Proceedings of 2024 International Conference on Guidance, Navigation and Control Volume 15
EditorsLiang Yan, Haibin Duan, Yimin Deng
PublisherSpringer Science and Business Media Deutschland GmbH
Pages48-57
Number of pages10
ISBN (Print)9789819622559
DOIs
Publication statusPublished - 2025
EventInternational Conference on Guidance, Navigation and Control, ICGNC 2024 - Changsha, China
Duration: 9 Aug 202411 Aug 2024

Publication series

NameLecture Notes in Electrical Engineering
Volume1351 LNEE
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Conference

ConferenceInternational Conference on Guidance, Navigation and Control, ICGNC 2024
Country/TerritoryChina
CityChangsha
Period9/08/2411/08/24

Keywords

  • deep reinforcement learning
  • motion planning
  • obstacle avoidance
  • space manipulator

Fingerprint

Dive into the research topics of 'Research on Obstacle Avoidance Motion Planning of Space Manipulator Based on Reinforcement Learning'. Together they form a unique fingerprint.

Cite this