Abstract
In the case that the high compliance and high safety are required, the use of pneumatic muscle for robot joints is a goodchoice. In this paper, a bio-mimic structure of pneumatic muscle robot arm is introduced and the differential pressureprinciple is applied to control the arm. In order to solve the practical problems that difficult regulation with pneumaticmuscle and the lower accuracy of position control, a proportional stiffness regulation method with easy implementingfeature is presented through theoretical analysis and experiments. On this basis, the position control of the joint withpneumatic muscle is implemented, for which the adaptive and self-learning neural-PSD controller is adopted. Testsindicate that the joint stiffness is easy to regulate by the method and joint operation shows a quick response and highaccuracy.
