On the Role of the Action Space in Robot Manipulation Learning and Sim-to-Real Transfer
This work addresses the problem of designing effective RL algorithms for robot manipulation by providing actionable recommendations for action space selection, though it is incremental in nature.
The study investigated how different action spaces affect robot manipulation learning and sim-to-real transfer, training over 250 RL agents across 13 control spaces in simulated tasks and evaluating real-world performance to identify good and bad characteristics.
We study the choice of action space in robot manipulation learning and sim-to-real transfer. We define metrics that assess the performance, and examine the emerging properties in the different action spaces. We train over 250 reinforcement learning~(RL) agents in simulated reaching and pushing tasks, using 13 different control spaces. The choice of spaces spans combinations of common action space design characteristics. We evaluate the training performance in simulation and the transfer to a real-world environment. We identify good and bad characteristics of robotic action spaces and make recommendations for future designs. Our findings have important implications for the design of RL algorithms for robot manipulation tasks, and highlight the need for careful consideration of action spaces when training and transferring RL agents for real-world robotics.