Independent stiffness and force control of antagonistic pneumatic artificial muscles joint

Abstract

Pneumatic artificial muscles (PAMs) get widely adopted due to their similar properties compared with biological muscles. PAM is a typical pneumatic actuator. Its capability to inflate is the main characteristic, and it can produce indiscriminate expansibility. The joint of this experiment is all composed of antagonistic PAMs. In order to gain a better performance, the experiment firstly established the model of PAMs. Concerning the fact that in essence, PAM is an elastic actuator, built its dynamic model, and got its stiffness model based on the assumption that PAM can be seen as variable stiffness springs. Its force and stiffness are proportional to the application pressure. This experiment requires an accurate mathematical model and a precise control method to achieve a good performance. But given that complex nonlinear dynamics exists in PMAs, an accurate mathematic model is hardly available, which leads us to consider the sliding control method. This suggestion guarantees independent controls on the force and stiffness of the joints. The experiment showed that the joints' control algorithm gave the celerity and accuracy on controlling the force and stiffness, and it compared results of different loads.