Our project aims to explore and analyze the specific performance task of a robot arm setting a chessboard with the RLBench environment. Our approach involves experimenting with existing algorithms in the RLBench environment, tweaking them to improve their performance, and conducting a comparative study to understand the impact of these modifications.
The robot arm in the RlBench environment looks like the image below:
The robot arm sets up a chessboard, shown in the figure below: