Optimization-driven design and experimental validation of a hydraulic robot leg mechanism

Abstract

Hydraulic-actuated legs for quadruped robots excel in producing high force and offering precise control. Although the overall efficiency of hydraulic servo systems can be limited by pump and valve losses, the local mechanical efficiency from the actuator to the leg mechanism can be relatively high. This study introduces an optimization driven methodology for designing and validating robotic leg mechanisms using evolutionary algorithms. By solving three distinct optimization problems, the study addresses trajectory tracking accuracy and force transmission efficiency. The resulting design was experimentally validated, demonstrating reliable motion reproduction with minimal deviation and achieving a force transmission efficiency of 94%. These results demonstrate the feasibility of translating optimization outcomes into high-performing physical prototypes, providing a robust framework for future robotic mechanism development.