ECoDe: A Sample-Efficient Method for Co-Design of Robotic Agents

TL;DR

ECoDe introduces a sample-efficient co-design method for robotic agents by leveraging multi-fidelity exploration and universal policy learning, enabling effective simultaneous optimization of design and control.

Contribution

The paper presents a novel multi-fidelity exploration strategy with universal policy learning to improve sample efficiency in robotic co-design.

Findings

Outperforms baseline methods in various design tasks.

Achieves significant reduction in data requirements.

Produces innovative and simplified robot designs.

Abstract

Co-designing autonomous robotic agents involves simultaneously optimizing the controller and physical design of the agent. Its inherent bi-level optimization formulation necessitates an outer loop design optimization driven by an inner loop control optimization. This can be challenging when the design space is large and each design evaluation involves a data-intensive reinforcement learning process for control optimization. To improve the sample efficiency of co-design, we propose a multi-fidelity-based exploration strategy in which we tie the controllers learned across the design spaces through a universal policy learner for warm-starting subsequent controller learning problems. Experiments performed on a wide range of agent design problems demonstrate the superiority of our method compared to baselines. Additionally, analysis of the optimized designs shows interesting design alterations, including design simplifications and non-intuitive alterations.