Stiffness control of bionic joint for a musculoskeletal leg mechanism

摘要

© 2016, World Scientific Publishing Co. Pte Ltd. All rights reserved. Pneumatic artificial muscles (PAMs) have properties similar to a biological muscle, and are widely used in robotics. A musculoskeletal leg mechanism driven by PAMs is presented in this paper for a quadruped robot to achieve jumping movement with relatively higher-speed. There is contact force between foot and environment of robot with high-speed. In order to reduce the impact and improve the movement performance, the joint stiffness is controlled in real-time with contact stiffness and motion speed so as to reduce the impact due to contact force between the robot foot and the environment at high-speeds and improve the movement performance. A synchronous control method for position and stiffness is used. PID and BP neural network controls are developed for the antagonistic knee joint. Experimental investigations, providing a comparative assessment of performance of the PID and neural network approaches are carried out. A close match is achieved between theoretical analysis and experimental results.