Research On Control Algorithms Of 3-DOF Joint Driven By Pneumatic Muscle Actuator

Pneumatic muscle actuator(PMA) is mainly made of rubber tube, braided fibrous shell and stiff connectors. When the rubber tube is pressurized, PMA shortens in axis and imposes contraction force on load. Compared with traditional actuators, PMA has the advantages of high power/volume and high power/weight ratios and flexibility similar to biological muscles which make it has a wide application prospect in the fields of bionic robot and medical rehabilitation.This paper mainly focuses on control algorithms of 3-DOF joint driven by PMA. The main work is as follows:First of all, the mechanical model of PMA is studied based on stiffness characteristic and BP neural network respectively. The former has a more clear expression than the latter, but its prediction precision is relatively low.Secondly, the kinematics and dynamic equations of the 3-DOF joint are derived and then verified by numerical simulation and Adams. According to above equations, control algorithms including computed torque method, robust control, sliding mode control and fuzzy-sliding mode control are designed.Finally, these algorithms are compared by combined simulation and experiment. Results show that robust control and fuzzy-sliding mode control can effectively compensate for model errors and external disturbance and have high control precision.