Research On Grasping Control Of Humanoid Arm Based On SEMG Signals

The surface electromyography(sEMG)signals can reflect the contraction of the muscles and can control the movements of the humanoid arm.The application of the sEMG signals on the human arm to control the grasping object of humanoid arm can effectively promote the application of humanoid arm in rehabilitation medical field.This paper studies from 4 aspects,arm movements sEMG signals processing,arm movements sEMG signals recognition,humanoid arm grasping control,humanoid arm grasping control experimental research and so on,in order to achieve humanoid arm grasping objects.The main work is as follows:(1)This paper makes some improvements to the methods of sEMG signals processing proposed by predecessors,and puts forward a set of methods to deal with the sEMG signals of the arm movements.First,the self assembled sEMG signals acquisition device is set up to collect the sEMG signals of the arm movements.Then,active segment extraction and denoising are performed on the collected arm movements sEMG signals.Finally,the eigenvalues of the sEMG signals are extracted and the dimensionality of the extracted eigenvalues are reduced in order to recognize the arm movements.(2)In this paper,the global search optimal solution principle of search optimization algorithm(SOA)and the ability of extreme learning machine(ELM)to deal with nonlinear relations were used,and propose a human arm movements recognition method based on searcher optimization algorithm to optimize the extreme learning machine(SOA-ELM).The ELM is trained by the processed sEMG signals,and the optimal input level weights and hidden layer node thresholds of ELM are searched by SOA to build the best ELM model.Finally,5 kinds of arm movements are identified by SOA optimized ELM model.The experimental results show that the SOA-ELM can effectively recognize 5 kinds of arm movements.(3)In this paper,the model recognition algorithm described above is used to recognize the arm movements,and the recognition results are used to control the movements of humanoid arm.The sEMG signals of the arm are used for quantitative recognition of the grasp strength of the hand,then,pressure sensors are used to perceive the grasp strength of a humanoid hand,and the recognition results and the perceptual results are applied to the grasping control of the humanoid hand based on the force feedback.Finally,the MATLAB and LabVIEW are used to design the movements control program of the humanoid arm and the grasping control program of the humanoid hand based on the force feedback.(4)In this paper,self assembled signal acquisition device,self assembled humanoid arm,computer and other hardware and sEMG signals acquisition program,human arm movements model training program,humanoid arm real-time grasping control program and other software are used to build an experimental platform for humanoid arm to grasp objects.The experimental platform is used to complete the grasping control experiment of human arm based on the sEMG signals.