Learning Task-Specific Dynamics to Improve Whole-Body Control

TL;DR

This paper introduces a method to learn task-specific dynamics for humanoid robots, reducing feedback gains and improving compliance and accuracy in whole-body control.

Contribution

It presents a novel approach to incorporate learned task-space reference accelerations, enhancing task execution and system compliance while reducing heuristic tuning.

Findings

Improved task execution accuracy in simulations and real-world tests.

Reduced feedback gains lead to more compliant robot behavior.

Effective learning of task-space dynamics of the robot's center of mass.

Abstract

In task-based inverse dynamics control, reference accelerations used to follow a desired plan can be broken down into feedforward and feedback trajectories. The feedback term accounts for tracking errors that are caused from inaccurate dynamic models or external disturbances. On underactuated, free-floating robots, such as humanoids, good tracking accuracy often necessitates high feedback gains, which leads to undesirable stiff behaviors. The magnitude of these gains is anyways often strongly limited by the control bandwidth. In this paper, we show how to reduce the required contribution of the feedback controller by incorporating learned task-space reference accelerations. Thus, we i) improve the execution of the given specific task, and ii) offer the means to reduce feedback gains, providing for greater compliance of the system. %With a systematic approach we also reduce heuristic tuning of the model parameters and feedback gains, often present in real-world experiments. In contrast to learning task-specific joint-torques, which might produce a similar effect but can lead to poor generalization, our approach directly learns the task-space dynamics of the center of mass of a humanoid robot. Simulated and real-world results on the lower part of the Sarcos Hermes humanoid robot demonstrate the applicability of the approach.