Task-Based Design and Policy Co-Optimization for Tendon-driven Underactuated Kinematic Chains

TL;DR

This paper presents a method for co-optimizing the design and control policies of tendon-driven underactuated manipulators using reinforcement learning, enabling effective transfer from simulation to real hardware.

Contribution

It introduces a general model for tendon-driven transmission and demonstrates a co-optimization approach for design and control of underactuated manipulators.

Findings

Optimized transmission and control policies transfer reliably to hardware.

Reinforcement learning effectively co-optimizes design and control.

The method improves manipulator performance in real-world tasks.

Abstract

Underactuated manipulators reduce the number of bulky motors, thereby enabling compact and mechanically robust designs. However, fewer actuators than joints means that the manipulator can only access a specific manifold within the joint space, which is particular to a given hardware configuration and can be low-dimensional and/or discontinuous. Determining an appropriate set of hardware parameters for this class of mechanisms, therefore, is difficult - even for traditional task-based co-optimization methods. In this paper, our goal is to implement a task-based design and policy co-optimization method for underactuated, tendon-driven manipulators. We first formulate a general model for an underactuated, tendon-driven transmission. We then use this model to co-optimize a three-link, two-actuator kinematic chain using reinforcement learning. We demonstrate that our optimized tendon transmission and control policy can be transferred reliably to physical hardware with real-world reaching experiments.