Review of Motion Control Research for Bionic Quadruped Robots: From Biological Inspiration to Algorithm Implementation

Abstract

Bionic quadruped robots offer significant potential for navigating complex and unstructured environments, yet their performance still lags notably behind that of natural quadrupeds in terms of efficiency and adaptability. This review systematically explores the progress and challenges in motion control technologies for biomimetic quadruped robots, drawing insights from biological models and robotic algorithms. It analyzes essential biological mechanisms such as the spring-mass model and tendon elasticity principles, alongside robotic control approaches including central pattern generators (CPGs) and reinforcement learning. A quantitative analysis of current robot performance indicates that existing systems exhibit limited obstacle negotiation abilities and substantial inefficiencies. The paper identifies core challenges in structural design rigidity, environmental adaptability, and computational efficiency. Recommendations for future research include integrating biologically inspired elasticity into structural designs, refining reinforcement learning algorithms to lower computational costs, and conducting extensive field testing in complex scenarios. Addressing these priorities promises significant advancements toward practical deployment in rescue, exploration, and similar demanding applications.