Stochastic Navigation And Control Of The Mobile Robot Based On Electroencephalogram Teleoperation

The "brain-computer control" is the direct control for the external devices by the human brain consciousness,so as to reduce or even replace operation by the limbs to achieve flexible control for the external devices,and the key technology is the brain-computer interface(BCI).BCI is a new way of the information interaction,and the core idea is to establish a direct communication channel between the brain and the external devices rather than normal output paths of peripheral nerves and muscles tissue.In the thesis,the steady-state visual evoked potential(SSVEP)and motor imagery(MI)spontaneous potential are analyzed and studied,and then stochastic navigation and control system of the mobile robot based on electroencephalogram(EEG)teleoperation is designed.The SSVEP-BCI system has higher transmission rate and recognition rate.We design the visual stimulation interface by analyzing the physiological basis of the SSVEP signals.By comparing different classification methods for the SSVEP signals,the thesis adopts the multivariate synchronization index(MSI)frequency identification algorithmis to decode the SSVEP signals,and the SSVEP-BCI system is constructed.For the MI spontaneous potential,EEG signals are filtered through a band in which event-related desynchronization(ERD)and event-related synchronization(ERS)physiological phenomenon is obvious.And the MI-BCI system is designed by combining common spatial pattern(CSP)feature extraction algorithm and support vector machine(SVM)method.Combining with the SSVEP-BCI system and MI-BCI system,the thesis designs the stochastic navigation and control system of the mobile robot based on EEG teleoperation.Considering the constraints of obstacles under the limited environment,the thesis breaks the traditional control method based on direction driven,and an innovative brain-controlled method based on probability potential fields(PPF)is proposed,where the alternative variation of EEG signal corresponds to the distribution of the obstacles and the probability potential field.Then,the change of potential field is applied to the mobile robot and the control input is produced to drive the robot to realize brain-controlled teleoperation stochastic navigation and control.Besides,the visual location method based on particle filter is designed by combining with the indoor landmark environment features and the readings of grating encoder,which assists the mobile robot positioning precisely and build the 2-D environmental map.For the designed system,five volunteers teleoperation control the mobile robot to complete stochastic navigation and obstacle avoidance experiment by the EEG signals in indoors dynamic environment with multiple obstacles,and the statistical experimental results have confirmed the feasibility and effectiveness of the system.