Iterative Learning Control on Carrying Robot

At present, most of the control methods of the lower extremity exoskeleton carrying robot require to calculate the system's dynamic equation and create the control quantity at each moment, which don't make full use of the repeatability feature of human movement and inevitably causes the delay on the dynamic response of the system. In this paper, utilizing the dynamic repeat characteristics of human walking, the time-variable quasi-Newton iterative learning controller of the lower extremity exoskeleton carrying system is designed to improve the control speed and the precision, where the human-machine interaction is considered. When the modeling error is large or the dynamics equation is not so accurate, the feedback loop can be added to improve the system's performance. In this paper, the corresponding quasi-Newton method based on feedback is constructed, which is used to control the carrying robot. Simulation results show the valid of the given method.