Research On The Control Method Of Prosthetic Arm Based On Inertial Sensors

In daily production and life,the number of patients with upper limb amputation is on the rise because of various factors.They have a lot of inconvenience in daily life.Wearing bionic prosthetic arm can make them have independent life ability.It is urgent to study the technology of robotic limbs.In order to improve the practicality,the prosthesis system needs to perceive the motion intention of human body in real time.However,the use of bioelectricity(electromyography or electroencephalogram)to perceive and identify the movement intention of human body is easily interfered by external or internal factors.Use for a long time can also affect the accuracy of recognition due to muscle fatigue or brain inattention.And the real time recognition is not fast because of the need to measure too many channels.Therefore,this paper was aimed to research on the control method of prosthetic arm based on inertial sensors.An inertial sensor was used to measure the posture of the shoulder joint of the wearer.In order to assist the wearer to complete the upper limb movement,an inertial sensor was adopted to measure the shoulder joint motion posture of the wearer for online human motion intention recognition,and the real-time trajectory control of the prosthetic arm was conducted according to the proposed shoulder-elbow and wrist mapping model.(1)Experimental design and attitude data acquisition.According to the degrees of freedom of shoulder joint and the spatial distribution of human body,six kinds of upper limb movements were designed,which are greeting,shaking hands,holding a cup,drinking water,blocking and elbow back.In this experiment,fifteen subjects were recruited to collect the attitude of the six movements under the standard experimental paradigm.Nine hundreds sets of experimental data were obtained for recognition for motion intention and establishment of mapping model between shoulder and elbow or wrist(2)Establish mapping model between the shoulder and elbow or wrist.Firstly,coordinate changes were carried out to transform the posture information of the upper limbs into joint motion information.Then the polynomial fitting method was used to fit curves representing the relationship between the shoulder and elbow or wrist.Finally the six mapping models of daily movements were established.(3)Propose the control method of prosthetic arm based on the inertial sensor.Firstly,propose a recognition method of upper limb motion intention based only on the shoulder joint.The average offline recognition rate of the above six actions was 85%.Then design the control system.A speed feedback control strategy was proposed to correct the deviation between the trajectory of the prosthetic target(model output)and the actual output trajectory in real time.The accuracy of the prosthesis movement was improved.(4)Experimental study on online control of prosthetic arm.The experimental platform of prosthetic arm was built,and the online control of prosthetic arm was carried out.The average online recognition rate of six typical daily movements reached 84%,and the real-time movement trajectory of the prosthetic arm deviated very little from that of the normal human body.The experimental results showed that the control method proposed in this paper can preliminarily realize the real-time control of the prosthetic arm,and verify the rationality and feasibility of the control method.