Task-Driven Co-Design of Mobile Manipulators
Raphael Schneider, Daniel Honerkamp, Tim Welschehold, Abhinav Valada
AI summary
Problem
Modular mobile manipulators typically mount arms in a tabletop configuration that restricts kinematics and limits performance for common mobile manipulation tasks requiring different workspaces and motions.
Approach
The authors introduce a concurrent design framework that optimizes arm mounting parameters by maximizing task success rates, using reinforcement learning for policy training in an inner loop and Bayesian Optimization (BOHB) for design exploration in an outer loop.
Key results
- First co-design method for modular mobile manipulator arm mounting
- Task-driven optimization significantly outperforms default tabletop mounting
- Direct task-based scoring generalizes better than heuristic manipulability metrics
- Optimized designs remain modular, affordable, and compatible with commercial components
Why it matters
Offers a practical, open-source methodology for developers and researchers to systematically improve modular robot designs for better real-world task performance.
Abstract
Recent interest in mobile manipulation has resulted in a wide range of new robot designs. A large family of these designs focuses on modular platforms that combine existing mobile bases with static manipulator arms. They combine these modules by mounting the arm in a tabletop configuration. However, the operating workspaces and heights for common mobile manipulation tasks, such as opening articulated objects, significantly differ from tabletop manipulation tasks. As a result, these standard arm mounting configurations can result in kinematics with restricted joint ranges and motions. To address these problems, we present the first Concurrent Design approach for mobile manipulators to optimize key arm-mounting parameters. Our approach directly targets task performance across representative household tasks by training a powerful multitask-capable reinforcement learning policy in an inner loop while optimizing over a distribution of design configurations guided by Bayesian Optimization and HyperBand (BOHB) in an outer loop. This results in novel designs that significantly improve performance across both seen and unseen test tasks, and outperform designs generated by heuristic-based performance indices that are cheaper to evaluate but only weakly correlated with the motions of interest. We evaluate the physical feasibility of the resulting designs and show that they are practical and remain modular, affordable, and compatible with existing commercial components. We open-source the approach and generated designs to facilitate further improvements of these platforms.