Home > Published Issues > 2021 > Volume 10, No. 10, October 2021 >

Comparison of Two Reinforcement Learning Algorithms for Robotic Pick and Place with Non-Visual Sensing

Muhammad Babar Imtiaz, Yuansong Qiao, and Brian Lee
Software Research Institute, Athlone Institute of Technology, Athlone, Ireland

Abstract—In this study, we perform a comparative analysis of two approaches we developed for learning to carryout pick and place operations on various objects moving on a conveyor belt in a non-visual environment, using proximity sensors. The problem under consideration is formulated as a Markov Decision Process. and solved by using Reinforcement Learning algorithms. Learning robotic manipulations using simple reward signals is still considered to be an unresolved problem. Our reinforcement learning algorithms are based on model-free off-policy training using Q-Learning and on-policy training using SARSA. Training and testing of both algorithms along with detailed a comparison analysis are performed in a simulation-based testbed. Our results prove our approaches to be successful in pick and place operations in non-visual industrial setups.

Index Terms—robotic manipulation, non-visual, Markov decision problem, reinforcement learning, Q-learning, SARSA

Cite: Muhammad Babar Imtiaz, Yuansong Qiao, and Brian Lee, "Comparison of Two Reinforcement Learning Algorithms for Robotic Pick and Place with Non-Visual Sensing," International Journal of Mechanical Engineering and Robotics Research, Vol. 10, No. 10, pp. 526-535, October 2021. DOI: 10.18178/ijmerr.10.10.526-535

Copyright © 2021 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.