Making use of design-aware policy optimization in legged-robotics co-design

Gabriele Fadini, Stelian Coros

Keywords: Robotics

TL;DR: This extended abstract explores the challenges of integrating robot optimization with learning to create adaptable design-aware policies. Issues and ongoing research directions are discussed.

Abstract: Our ongoing research aims to investigate the potential of integrating robot design optimization with reinforcement learning (RL). In co-design literature, exploiting the ties between design and control seems to be the key to unlocking otherwise unreachable performance. However, the problem of obtaining policies that will adapt to a range of different robots is still open. In this extended abstract, we would like to reason about the challenges that the policy optimization problem in this setting brings. Moreover, we hint at a few possible future research directions that may help in advancing robot morphology and design-aware control policies.

Submission Number: 8