Cycloidal Quasi-Direct Drive Actuator Designs with Learning-based Torque Estimation for Legged Robotics

2024-10-22

arXiv.org

AI Summary

Legged robots face concurrent challenges in achieving high torque density, lightweight design, and minimizing the sim-to-real gap under highly dynamic loading conditions. To address these, this work introduces a novel Cycloidal Quasi-Direct Drive (CQDD) actuator, the first to deeply integrate high-torque-density cycloidal gearing into a quasi-direct-drive architecture. Furthermore, we propose a physics-informed Actuator Network that enables data-driven, high-accuracy joint torque estimation. Our method reduces torque estimation error by 62% and improves the real-world deployment success rate of reinforcement learning policies by 3.1×. Experimental evaluation confirms that the CQDD actuator delivers exceptional mechanical robustness, high torque density (≥150 N·m/kg), and superior dynamic response (bandwidth > 30 Hz), enabling more agile and adaptive dynamic locomotion on legged platforms.

Technology Category

• Intelligent Robots: State Estimation
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Agentic search
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

This paper presents a novel approach through the design and implementation of Cycloidal Quasi-Direct Drive actuators for legged robotics. The cycloidal gear mechanism, with its inherent high torque density and mechanical robustness, offers significant advantages over conventional designs. By integrating cycloidal gears into the Quasi-Direct Drive framework, we aim to enhance the performance of legged robots, particularly in tasks demanding high torque and dynamic loads, while still keeping them lightweight. Additionally, we develop a torque estimation framework for the actuator using an Actuator Network, which effectively reduces the sim-to-real gap introduced by the cycloidal drive's complex dynamics. This integration is crucial for capturing the complex dynamics of a cycloidal drive, which contributes to improved learning efficiency, agility, and adaptability for reinforcement learning.

Problem

• Designing Cycloidal Quasi-Direct Drive actuators for legged robotics.
• Enhancing torque and dynamic load performance in lightweight robots.
• Developing a torque estimation framework to reduce sim-to-real gap.

Innovation

• Cycloidal gear mechanism for high torque density
• Quasi-Direct Drive framework for lightweight robotics
• Actuator Network for torque estimation and sim-to-real gap reduction