Prescribed performance control of a link-type exoskeleton powered by pneumatic muscles with virtual elasticity

This paper presents a rehabilitation exoskeleton featuring a link-type design, driven by pneumatic muscles (PMs), with the driving torques transmitted to the leg orthoses via multiple linkages. The inherent complexity of this link-type exoskeleton presents significant challenges in both modeling and control. To address these challenges, we construct constrained kinematics and dynamics through geometric analysis, and introduce a robust prescribed performance controller with a virtual elastic element (PPC-VE) to enhance the safety of passive gait training. The controller has the capability to directly adjust the intensity of chattering in control inputs and ensure system transient/steady states performance by introducing a non-zero proxy mass and an error transformation. Theoretical analysis indicated the uniform ultimate boundedness of the states, with varying behaviors observed in the controller as the proxy mass changed. Simulation and experimental results demonstrated the validity of the dynamic model, and the proposed controller effectively enhanced system safety by addressing output constraints and reducing abrupt variations in control inputs.