Research On The Control System Of Three-degree-of-freedom Electromyographic Prosthesis (upper Limb) Based On PSoC

At present,the number of upper limb disabled people is increasing year by year,and the traditional prosthesis with inconvenient use and poor bionic effect can no longer meet the life and work needs of armless handicapped.With the development of science and technology,a kind of intelligent prosthesis which is controlled by sEMG has become a research hotspot.As artificial limbs based on threshold control method are not intuitive and flexible on market.Pattern-recognized controllors is still in the research stage,and most of them are still in the experimental simulation state,among them,the mode recognition controller with better control effect also has the disadvantages of poor portability and practicality,which is far from the level of people's expectations.In this paper,In order to achieve the practical application of three-degree-of-freedom myoelectric prosthesis.A portable prosthetic control system based on PSoC and DP-PSO-SVM is proposed.The system is built using software and hardware collaborative design methods,the hardware part of the system mainly includes:PSoC main control circuit,portable sEMG signal acquisition circuit,power supply circuit and prosthetic drive circuit;The software part of the system includes:AD acquisition of EMG signals,?R digital filtering,motion start and end point judgment,feature value extraction and pattern recognition procedures.In the system design research,in order to improve the portability of the system,when designing the sEMG signal acquisition hardware circuit,we use the software and hardware coordination method to integrate the acquisition circuit and the control circuit based on PSoC.Then the portability can be easily installed inside the prosthesis.the system integration is improved in this way.Furthermore,in order to improve the recognition accuracy of pattern recognition,we adapt the supervised learning algorithm-support vector machine,and use the dual-population particle swarm optimization algorithm(DP-PSO)to solve the problem that the recognition rate of some actions is too low,and the overall recognition rate of the system is improed.The control system uses PSoC as the main controller,and it uses the portable acquisition circuit to collect the four-channel sEMG signal of the arm.Then,we deal with the signals wtih digital filtering,start and end point judgment and eigenvalue extraction inside the PSoC.The component feature vector is uploaded to the upper computer to train the DP-PSO-SVM classifier mathematical model.After the model parameters were written into the PSoC,the pattern recognition of the arm movement can be completed without the computer.Finally,the pattern recognition control of the six kinds of motions of the three free holiday limbs can be realized by the driving circuit,these motions include open hands,close hands,internal rotation of the wrist,external rotation of the wrist,elbow extension and elbow flexion.In order to verify the effect of the control system,we design offline simulation comparison and online control prosthetic experiment.The experimental results show:the classifier using DP-PSO-SVM algorithm has a recognition rate of 96.1%,which increases the recognition rate by 4%compared with the classifier based on PSO-SVM algorithm;the controller has an online recognition rate of 95.8%for 6 actions,It has a high recognition rate of motion and meets the requirements of real-time.It satisfies the control requirements of the three free holiday limbs,and it greatly improves the practicality of the prosthesis.