Force sensorless compliance control of a lower-limb exoskeleton robot

Lower limb rehabilitation machines are widely used to enhance the mobility function of elderly people and patients suffering from spinal cord injury and stroke. In this paper, a four degrees of freedom (DOF) lower body exoskeleton with a model-based compensation control framework is proposed to support hip-knee rehabilitation. The exoskeleton control movement is realized by designing a trajectory for each leg movement. A function approximation technique (FAT) based adaptive control is applied to each two DOF legs during a rehabilitation task. Using the FAT based adaptive control, the natural system dynamic is adaptively compensated for without knowing the acceleration feedback and system dynamics. Simulation results show the disturbance observer successfully estimated the user's applied leg force. The leg force is given as the input to compliance control applied in 2 DOF motors when performing a walking sequence on a treadmill. Moreover, the FAT-based adaptive control outperformed the standard proportional-differential (PD) control and could accommodate different subjects without any changes to control parameters.