Research On Continuous Shared Control System Of Mobile Robot Based On Motor Imagery EEG

Brain Computer Interface(BCI)technology realizes direct interaction between human brain and external devices by decoding brain activity signals.The BCI system based on motor imagery(MI)allows users to spontaneously imagine limb movements to output control intentions and complete the control of devices such as robotic arms and mobile robots.However,in this control process,it is often difficult to effectively combine the user’s intention and the robot’s autonomous control,and the user’s control process of the system in an unknown environment also has unnatural limitations.This thesis focuses on the above issues,proposes a continuous shared control(SC)system for mobile robots based on motor imagery electroencephalography(EEG).For the navigation task of unknown environment,the mobile robot system firstly performs simultaneous localization and mapping(SLAM)based on the visual method,and then performs path planning through grid maps to generate autonomous control signals.In this process,the user achieves the continuous control output on the velocity of the mobile robot by imagining the hand movement,then the whole system combines the EEG control instructions and the autonomous navigation control instructions to complete the continuous shared control.For the continuous control based on motor imagery EEG,this thesis proposes a regression-based EEG continuous control method.The amplitude features of EEG signals are extracted using an Autoregressive(AR)model,and online continuous decoding is performed using linear regression.The offline training experiment without feedback and online virtual cursor control experiment with feedback are used to analyze the training situation of the motor imagery EEG paradigm and the continuous control ability of the subjects.For the shared control of mobile robots based on EEG and autonomous navigation,this thesis proposes a shared control method on the velocity of mobile robots.The method control the robot’s linear velocity by fusing the EEG control signal and the autonomous navigation control signal through a sharing coefficient which is dynamically adjusted with the change of the environment.Then the local path planner combines the shared control signal to sample and evaluate in the velocity space to output control signals.The online shared control experiment and online autonomous navigation experiment are used to analyze the environment mapping performance,the completion of navigation tasks and the dynamic change of the sharing coefficient of the mobile robot shared control system.In this thesis,the proposed system uses the shared control method to integrate the motor imagery EEG control with the autonomous control of the robot,which improves the performance of the mobile robot to move in an unknown environment and reduces the user’s mental burden.The feasibility of the whole system was verified by recruiting subjects to conduct experiments.