Coordinated Motion Planning of Dual-arm Space Robot with Deep Reinforcement Learning

In this paper, we focus on coordinated motion planning of dual-arm robot. The kinematics model of the robotic arm is established by Denavit-Hartenberg (D-H) coordinate method and the mathematical model of the cooperative motion planning problem is established. The rapidly-exploring random trees (RRT...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:2019 IEEE International Conference on Unmanned Systems (ICUS) s. 469 - 473
Hlavní autori: Tang, Mengying, Yue, Xiaofei, Zuo, Zhan, Huang, Xiaoping, Liu, Yanfang, Qi, Naiming
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 01.10.2019
Predmet:
Collision avoidance
collision detection
deep deterministic policy gradient algorithm
dual-arm robot
Kinematics
Manipulators
Mathematical model
motion planning
Planning
rapidly-exploring random trees algorithm
Robot kinematics
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:In this paper, we focus on coordinated motion planning of dual-arm robot. The kinematics model of the robotic arm is established by Denavit-Hartenberg (D-H) coordinate method and the mathematical model of the cooperative motion planning problem is established. The rapidly-exploring random trees (RRT) algorithm and the deep deterministic policy gradient (DDPG) algorithm are used to carry out dual-arm coordinated motion planning, respectively. The simulation results show that these algorithms can effectively complete the robot arm motion planning task, but the RRT improved algorithm cannot balance the planning efficiency and result optimization. Compared with the RRT algorithm, the DDPG algorithm trains the model through continuous trial and error to optimize its planning strategy. The trained model can be used to obtain an optimized path and it can ensure the efficiency of the planning with the optimized strategy.
DOI:10.1109/ICUS48101.2019.8996069