Research On Robot Control Method Based On EMG Signal

The bioelectric signals of human body include EMG signal,EEG signal,ECG signal and EOG signal.These bioelectric signals contain a great deal of information about human behavior and are the direct response to the intention of human motion.With the rapid development of human-robot interaction technology in the field of robotics,it is particularly important to realize the similar interaction between human beings and robots.As the biological signal of the human body is a direct response to the intention of human behavior,it has also become a research hotspot in the field of human-robot interaction.Compared with several other kinds of human bioelectrical signals,surface EMG signal is more convenient to collect and process.This paper focuses on the human-robot interaction technology between human and KUKA based on the human body's surface EMG signal.In this paper,the generation mechanism of surface EMG is first analyzed and the surface EMG signals of two muscles in the forearm of the human arm are acquired by the Trigno EMG wireless acquisition equipment of Delsys Company in the United States.Feature extraction is the key to the follow-up processing of surface EMG signal.Because human-robot interaction requires high real-time EMG signal recognition,we focus on extracting the time-domain features of EMG signals and then Different feature sets were combined into eigenvectors,and three different gesture actions were classified by Linear Discriminant Analysis(Linear Discriminant Analysis,LDA)classification model respectively.The effects of different EMG feature combinations on the classification good and bad were studied.Select the best eigenvalues and combine them into eigenvectors as sample data for training or testing,and experimentally achieve good gesture recognition results.Secondly,the 7-DOF KUKA robot in the laboratory is analyzed structurally,the kinematics is modeled by the classical Denavit-Hartenberg(D-H)method,the positive kinematics and the inverse kinematics are studied,Inverse kinematics as an important basis for the study of trajectory planning,this paper uses Jacobi than pseudo-inverse iterative method.Then the trajectory planning movement of the manipulator in joint space and Cartesian space is studied.The study of point-to-point total joint motion and single-joint reciprocating motion was carried out in the joint space.The single-joint reciprocating motion could start and stop the manipulator at any moment.Point-to-point linear trajectory planning in Cartesian space,reciprocation in Cartesian directions,and unidirectional motion in Cartesian XYZ plus and minus direction,where the latter two can achieve the start and still at any moment.Finally,the electromyography acquisition system is integrated with the KUKA robot system and the end effector claw to form the platform of the human-robot interaction experiment.The electromyographic control based on discrete recognition is used to control the KUKA machine online and in real time Arm trajectory movement,the end of the actuator claw hand open and closed by the independent control of the serial communication matlab,together to complete the capture and placement of the bottle task to achieve the application of human-robot interaction technology.