Study Of Embedded Control System For A Pneumatic Power-Assist Exoskeleton Based On DSP

Exoskeleton robot is a wearable device,which can follow the human movement and simultaneously provide power for the human body.In the medical treatment filed,exoskeleton robots can be used for rehabilitation therapy and assist people with motor impairment or disabilities to walk.In military and industry filed,it can provide power for wearers,increase load and reduce physical consumption.One of the key technologies of exoskeleton robots is to provide assistance to better follow the wearer's movement,quickly identify human motion intentions and control exoskeleton movement.And the control system,as the control center of exoskeleton robot,determines the performance of exoskeleton robot.In this paper,an exoskeleton robot control system based on embedded scheme for the upper extremity exoskeleton robot is studied.Firstly,the embedded control system based on single-chip DSP processor is proposed.The hardware design of signal conditioning circuit,high-speed on-off valve driving circuit and signal acquisition circuit are completed to meet the needs of miniaturization,low power consumption and portability.Besides,the embedded system software is developed and the upper computer measurement and control system for testing and experiment is completed.Secondly,aiming at the need of human-machine cooperative motion control of upper extremity exoskeleton robot,the research on human upper extremity motion intention recognition algorithm,human motion state detection algorithm based on logistic regression and the pneumatic muscle driving control algorithm based on gas pressure controller is conducted based on the collected pneumatic muscle internal pressure signal,joint angular displacement signal and human surface electromyogram signal etc.is carried out,providing algorithm support for the implementation of embedded control system.On this basis,the effects of PID and sliding mode variable structure control algorithm on pneumatic muscle pressure following control are compared through experiment.The results show that sliding mode variable structure control has better tracking accuracy than PID control in different periods and loads.Thirdly,the foregoing algorithms of human upper limb motion intention recognition,motion state detection and pneumatic muscle drive control are programmed and implemented in embedded control system,and the prototype of the system is completed.Based on the system prototype,experiments are conducted to verify the assisting effect of the control system in the actual motion process.The effectiveness of the proposed algorithm in human-machine cooperative motion control is verified by comparing the assisting effects under different working conditions,including different loads,different power ratios and continuous motion with motion and stop switching.The results show that the above algorithm can better recognize the human upper limb movement intention and accurately detect the current motion state,and provide timely assistance for human upper limb movement.Finally,according to the full-text research and experimental results,this paper summarizes the paper and puts forward the prospects for the follow-up work.