Interactive Imitation Learning of Bimanual Movement Primitives

TL;DR

This paper introduces SIMPLe, a novel algorithm for learning and correcting bimanual robotic movements from human demonstrations, ensuring synchronization and adaptability through Gaussian Process Regression and uncertainty-based stiffness regulation.

Contribution

The paper presents a new interactive learning algorithm for bimanual tasks that guarantees convergence and allows motion reshaping via human feedback and external perturbations.

Findings

Successfully synchronized dual-arm movements using kinesthetic feedback.

Reshaped bimanual motions to adapt to new task parameters.

Demonstrated convergence of motions close to demonstrated trajectories.

Abstract

Performing bimanual tasks with dual robotic setups can drastically increase the impact on industrial and daily life applications. However, performing a bimanual task brings many challenges, like synchronization and coordination of the single-arm policies. This article proposes the Safe, Interactive Movement Primitives Learning (SIMPLe) algorithm, to teach and correct single or dual arm impedance policies directly from human kinesthetic demonstrations. Moreover, it proposes a novel graph encoding of the policy based on Gaussian Process Regression (GPR) where the single-arm motion is guaranteed to converge close to the trajectory and then towards the demonstrated goal. Regulation of the robot stiffness according to the epistemic uncertainty of the policy allows for easily reshaping the motion with human feedback and/or adapting to external perturbations. We tested the SIMPLe algorithm on a real dual-arm setup where the teacher gave separate single-arm demonstrations and then successfully synchronized them only using kinesthetic feedback or where the original bimanual demonstration was locally reshaped to pick a box at a different height.