Real Robot Challenge Stage 1: Cartesian Position Control with Triangular Grasp and Trajectory Interpolation

Anonymous

Keywords: Dexterous Manipulation, Cartesian Position Control

Abstract: We present an extension to our runner-up approach from the previous edition of the Real Robot Challenge. To solve the task of sequential goal reaching we focus on two aspects to achieving near-optimal trajectory: Grasp stability and Controller performance. In the simulated challenge, our method relied on a hand-designed Pinch grasp combined with Trajectory Interpolation for better stability during the motion for fast goal reaching. In Stage 1, we observe reverting to Triangular grasp to provide a more stable grasp when combined with Trajectory interpolation, possibly due to the sim2real gap. In future stages, we propose using sampling-based planning methods and multi-agent pathfinding algorithms for complex rearrangement tasks. The video demonstrations for our approach our available at https://youtu.be/dlOueoaRWrM.