Controller design and experimental verification of a self-optimizing motion controller for smart exercise machines

Control of single degree of freedom concentric exercise machines is considered. An adaptive resistance controller is designed under the restriction that the resistance mechanism has only a braking capability. Passivity and bounded tracking error of the closed loop system are proven, and an optimal exercise protocol is proposed based on the velocity dependence of muscular force. The controller is implemented on two exercise machine platforms, an arm exercise machine and a recumbent bicycle. Identification of the nonlinear system dynamics of the exercise motions and torque output of the resistance mechanisms is performed. Clinical studies are presented that verify the optimality of the proposed exercise criteria both in the short term and when fatigue is accounted for.