Study On Fatigue Mechanism Of SSVEP And Its Application In Brain-controlled Wheelchair System

Brain-computer interface(BCI)system is a new technology which can directly construct information pathways between the human brain and the controlled device.It provides a new means for giving instructions to the external environment.At present,the research of brain-controlled wheelchair has attracted more and more attention.The elderly people and the patients with movement disorders can control wheelchair by mind independently instead of others' help,so as to improve their self-care ability.However,the appearance of fatigue symptoms during long time operation deteriorates the control performance.Currently,the studies of the fatigue mechanism based on EEG are still in the infancy stage,and it remains a significant challenge for the traditional methods to investigate the fatigue mechanism from multi-channel EEG data.Focusing on the brain-controlled wheelchair system,this dissertation chooses SSVEP paradigm which has the advantages of relatively high rate of information transmission and low training demand.This dissertation elaborately designs the SSVEP-based experiment to record EEG signals in normal and fatigue status,and then classifies the normal and fatigue SSVEP signals and study the effect of fatigue behavior on classification accuracy.The results indicate that classification accuracy declines in condition of fatigue.Simultaneously,we find SVM has a better classification performance.The average accuracy under normal and fatigue status is 92.67% and 72.91% respectively.In order to further study the fatigue mechanism,this dissertation constructs modality transition-based networks from four channels EEG signals which are located near the brain occipital region,and then calculate the "small-world-ness" index of the derived networks.The results suggest that the small-world-ness in the fatigue status is significantly enhanced.These interesting findings suggest that,the brain network tends to be more clustered locally and more integrated globally so as to complete the cognitive tasks in the fatigue status.In order to realize the real-time control of the wheelchair,this dissertation designs the SSVEP-based brain-controlled wheelchair system and carries out online obstacle avoidance experiment.Considering the timeliness of the brain-controlled wheelchair,most of the existing EEG acquisition devices restrict the number of acquisition channels and the portability of equipment,which have reduced the timeliness of the system.Therefore,the design of portable acquisition equipment has a broad prospect.In the optimization research of the acquisition equipment,the key is how to establish data transmission between portable acquisition equipment and wheelchair upper computer.This dissertation designs a portable EEG acquisition communication module for brain-controlled wheelchair system.We choose USB for communication which has a higher transmission speed than common UART serial port.In addition,it has a good anti-interference ability compared with wireless communication and can support brain-controlled wheelchair system to maintain stable data communication in complex environments.Subjects generate brain-controlled EEG signals by looking at the SSVEP motivational interface,which can be transmitted by the designed communication module.The instructions can be obtained by analyzing EEG data online,which allows completing the obstacle avoidance task.