Bioelectrical Signal Analysis And Its Application In Motion Control

With the development of human-computer cooperation and intelligence,bioelectrical signals such as brain electricity,myoelectricity,etc.are applied to complex motion control.At present,the research and development of prosthetic hands based on surface electromyography signals has become a hot research topic at home and abroad.In order to solve the problems of insufficient channels of existing s EMG signal acquisition equipment,low control strength of prosthetic hand and low grasping accuracy of objects,a multi-channel s EMG signal acquisition system is designed in this paper.Meanwhile,gesture classification and fingertip force prediction algorithm are optimized,and the design of EMG motion control system is completed.The main work of this paper is as follows:(1)A gesture classification method based on support vector machine and fingertip force prediction method based on long-term memory network are proposed.Aiming at the problem of large amount of computation of kernel function and penalty factor optimal solution of support vector machine,genetic algorithm is used to optimize fitness function to reduce the complexity of solution and improve the real-time and accuracy of gesture classification;aiming at the problem of low grasping accuracy of EMG controlled prosthetic hand,longterm and short-term memory network method is used to improve the force feedback mechanism and improve the accuracy of fingertip force prediction through EMG recognition Degree.(2)A multi-channel acquisition scheme based on high-precision analog front end and cascade structure is designed.Firstly,according to the demand of high-precision surface electromyography signal acquisition,a surface electromyography signal acquisition scheme based on multi-channel differential and adjustable acquisition frequency is designed to achieve high-quality and multi-dimensional original surface electromyography signal acquisition;then,according to the demand of multi-channel acquisition,the multi-level parallel mode is adopted to realize the extended application of acquisition channels,which can meet the high-quality requirements such as 16 channels and 32 channels Then,according to the requirements of synchronous acquisition,the clock synchronization method strategy is proposed to achieve multi-channel high-speed and low delay synchronization effect;finally,according to the design requirements of portable,low power consumption and small volume,multi-level networking wireless communication technology is adopted to realize multi module data access.(3)Complete the design and analysis of the experimental environment platform.Firstly,aiming at the gesture classification verification experiment,the multi-dimensional dynamic bionic manipulator is constructed by using MATLAB Simulink motion simulation platform to realize the training environment of EMG recognition classification effect and improve the efficiency and convenience of the experiment;then,aiming at the fingertip force prediction verification experiment,combined with gesture motion control,the fingertip grasping experiment scheme based on piezoelectric film pressure feedback system is constructed.Finally,an experimental study on gesture classification and fingertip force prediction is carried out.The experimental results show that the accuracy of gesture recognition based on genetic algorithm is 93%,and the prediction error of fingertip force based on long-term and short-term memory network is less than 1.3N,which provides a new research idea for related experimental research.