Research For Wearable Robot Conctroller Based On SEMG Technology

With the rapid development of automated intelligent equipment,the application of robots and robotic arms is becoming more and more important in industrial production and daily life.Naturally,the demand for human-computer interaction systems for controlling robotic arms is increasing.Gesture action contains rich expression information,easy to use and intuitive and natural.This paper combines surface electromyography technology and pattern recognition algorithm to classify gestures,map different gestures to different robotic arms motions,and finally control the robotic arm motions through gestures.Using surface electromyography and flexible array electrode technology,a wearable surface electromyography acquisition system was designed.The support vector machine(SVM)algorithm was used in the PC recognition software to identify gestures in specific gestures.action.Finally,the 6 degree of freedom PUMA560 robotic arm motion is controlled by the corresponding gesture.The main research work of the thesis focuses on the following points:Firstly,the surface electromyography acquisition equipment was designed and fabricated.According to the requirements of surface electromyography collector,the corresponding hardware and software were designed,and a flexible acquisition electrode was designed and fabricated.The hardware part is designed and manufactured based on STM32 controller module and acquisition terminal module based on ADS 1298 chip.Completed circuit electronic design,wearable module design and overall system debugging.Secondly,this system uses the support vector machine(SVM)algorithm to realize gesture recognition.The host comupter software was written to filter out noise,to do feature extraction,classification and other processing of the original surface EMG signal.Compared related algorithm implementations,such as KNN,neural network algorithms,etc.Finally,using the combination of supporting camera algorithm and tree structure,the recognition of the gesture signal recognition is realized.And verify the practicability of the overall system through experiments.Finally,the virtual model and mapping relationship of the manipulator are established.The paper uses the Robotics Toolbox in MATLAB to model the robotic arm,which is the PUMA560 robotic arm.The parameters of the manipulator are constructed by standard D-H parameters,and the kinematics forward solution and kinematics inverse solution are used to estimate the joint angle parameters that need to be input for the manipulator motion.By calculating the gesture action to the motion parameter,the gesture set can map the three-axis motion of the robot arm to realize the final process of the gesture action to the control.