Research Of Continuous Control For Robot System Based On Motor Imagery EEG

Brain-computer interface(BCI)aims at providing a bridge between human brain and external devices.It is a communication control system that does not depend on the normal output channels of the peripheral nerves of the brain and muscles.The human brain electroencephalogram(EEG)signals are collected and extracted,and finally transformed into control signals.These control signals are used to control the external devices directly,this can help patients,who lost their ability to freely move and control their muscles because of severe neuromuscular diseases or motor system injuries,assisting them in their daily work so that they can reduce dependence on others.There are many types of EEG signals applied in the research of brain-computer interface.The motor imagery EEG signals are generated when subjects imagine the movement of the limbs,other than physical movement or reliance on external stimulation devices.A robot continuous control system based on motor imagery EEG is built,where the subjects adjust their EEG rhythm through motor imagery to control the robotic arm to complete the task.At present,the research on motor imagery EEG is mainly focused on the feature extraction and classification methods for binary or four classification problems of motor imagery EEG.In addition,most robot control systems based on motor imagery EEG are mainly discrete control systems.Each time a subject performs a motor imagery task,a control command is sent to control the robot to move along a predetermined trajectory.Research on two-dimensional control and two-dimensional continuous control based on motor imagery EEG is not deep enough,however the two-dimensional continuous control is the basis of realizing three-dimensional multi-degree-of-freedom control.Therefore,a two-dimensional continuous control system of robot based on motor imagery EEG is designed.When the subjects execute motor imagery task,the control signals are sent to the robot,in which both horizontal and vertical control signals are included,controlling the robot to move in a two-dimensional plane.The system includes a brain-computer interface module,a robot control module and a task state feedback module,forming a closed-loop control.In order to improve the two-dimensional control ability of subjects,a series of virtual cursor control training tasks with increasing difficulty were designed.Through these training tasks,subjects have the ability to regulate their EEG Mu/Beta rhythmic energy.Then,the cursor control and robot control are synchronized.Finally,several online experiments on robot continuous control based on motor imagery EEG were carried out.According to the experimental results,the BCI system allows the subject to control the robotic arm accurately through their "thoughts" to complete the grasp task.It also shows that trained subjects are able to maintain their ability to regulate their brain's Mu/Beta rhythm energy.