| Brain-Computer Interface(BCI)is the communication bridge between the human brain and external devices.This technology directly controls external devices by collecting human brain signals and converting them into control commands.Brain-controlled mobile robot is an important research direction of BCI.However,EEG signals are very weak and easily interfered,resulting in low recognition accuracy and information transmission rate.If only one type of EEG signal is used as a control command for a mobile robot,the control effect is often poor.This paper focuses on the research on the single-mode motor imagery(Motor Imagery,MI)brain-computer interface,proposes a control method combining motor imagery and blinking,establishes a hybrid brain-computer interface(HBCI)system,and applies it to the control of mobile robots.The research focus of the thesis is based on the following points:1.Propose two blink control methods for hybrid brain-computer interfaces.The method based on the count of blinks first uses t he active shape model algorithm to recognize the face,then calculates the real-time eye aspect ratio,and then counts the blink counts through a sliding time window.The method based on eye open and close state recognition firstly realizes the pupil locat ion by Adaboost algorithm and the ellipse fitting of pupil,and then judges the open and close state of the left eye/right eye by the number of melanin in the pupil.Finally,the number of blinks or the open and closed state of the eyes is converted into a control command,which is used to control the mobile robot to go straight,stop or start to turn.The results of eye recognition experiments show that the two blink control methods have higher recognition accuracy.2.Aiming at the problem of shared cont rol in the hybrid brain-computer interface system,in this thesis,a Petri net model is constructed describing the real-time operating state of the hybrid brain-computer interface system.According to the control strategy,Petri net coordinates eye control commands,EEG control commands and automatic obstacle avoidance commands to realize the human-machine coordinated control of mobile robots.3.Combining the eye blinking control system,the MI BCI,the Amigo Bot mobile robot system and the Petri net,a HBCI system based on eye blinking and motor imagery control is established.The HBCI system realizes the eye blinking control of the mobile robot to go straight,stop or start to turn,and MI control of the mobile robot to turn left or right.The online experiments further compare the effects of the two eye blinking methods proposed in this thesis to control the mobile robot.The experimental results show that the method based on the number of eye blinks has a wider application range,and the method based on eye open and close state recognition has better real-time performance.Compared with the single-modal BCI system for MI,the proposed HBCI system combining eye blink and motor imagery not only increases the number of control commands,but also improves the control effect of mobile robot,which is expected to improve the practicability of the BCI system. |
