Implementing Robotic Pick and Place with Non-visual Sensing Using Reinforcement Learning

In this study, we focus on learning and carrying out 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. Learning robotic manipulations using simple reward signals is still considered to be an unresolved problem. Our reinforcement learning algorithm is based on model-free off-policy training using Q-Learning. Training and testing are performed in a simulation-based testbed, proving our approach to be successful in pick and place operations in non-visual industrial setups.