Hardware Implementation of a Zero-Prior-Knowledge Approach to Lifelong Learning in Kinematic Control of Tendon-Driven Quadrupeds

TL;DR

This paper presents a bio-inspired, hardware-implemented zero-prior-knowledge learning algorithm for tendon-driven quadruped robots, enabling rapid, adaptive control of complex movements with minimal training time.

Contribution

It introduces a novel hardware-in-the-loop approach applying the G2P algorithm to achieve quick, autonomous control in tendon-driven quadrupeds without prior knowledge.

Findings

Robot learns cyclical movements in minutes

Achieves adaptive control with minimal trials

Demonstrates hardware implementation feasibility

Abstract

Like mammals, robots must rapidly learn to control their bodies and interact with their environment despite incomplete knowledge of their body structure and surroundings. They must also adapt to continuous changes in both. This work presents a bio-inspired learning algorithm, General-to-Particular (G2P), applied to a tendon-driven quadruped robotic system developed and fabricated in-house. Our quadruped robot undergoes an initial five-minute phase of generalized motor babbling, followed by 15 refinement trials (each lasting 20 seconds) to achieve specific cyclical movements. This process mirrors the exploration-exploitation paradigm observed in mammals. With each refinement, the robot progressively improves upon its initial "good enough" solution. Our results serve as a proof-of-concept, demonstrating the hardware-in-the-loop system's ability to learn the control of a tendon-driven quadruped with redundancies in just a few minutes to achieve functional and adaptive cyclical non-convex movements. By advancing autonomous control in robotic locomotion, our approach paves the way for robots capable of dynamically adjusting to new environments, ensuring sustained adaptability and performance.