The brain computer interface(BCI)is a technology that uses a special device tocollect EEG signals generated when different thinking activities of the brain are inprogress,and the features of EEG signals are extracted through the signal processingmeans and then translated into control commands to implement the directcommunication and control between the human brain and the external environment.The BCI technology has broad application prospects.In order to achieve the goal that uses EEG to control a smart car to forward,backward, turn left and turn right in real time, this paper focuses on the key technicalproblems of the real-time BCI system based on the left-right hand and left-right footmotor imagery and the key engineering and technical issues’ solutions, such as thesystem platform interface design and the data transmission associated with the BCI,and establishes a real-time BCI system based on four classification of motor imagery.The main work includes:1) Design a motor imagery experimental paradigm, and in order to avoidbringing the subject the emotional problems, which will lead to lowersignal quality,as a result of long experiment, single experiment is shortened to five seconds,including three seconds for imagery and two seconds for adjustment.21channelselectrodes related to the motor ribbon are selected to resolve the different electrodepositions on different subjects, and this will increase the versatility of the system. Atthe same time, it won’t increase the difficulty of the data processing because of thelarge amount of data.2) Use multi-Threading technology to design the real-time signal processingsystem to avoid the resource contention issues due to the excessive handling tasks,and it will improve the real-time system.Use Socket based on TCP/IP protocolbetween the acquisition system and the signal processing platform to achieve thereal-time data transmission, and this mean can improve expansibility of the system; The way of shared memory is used among the internal modules of the signalprocessing system to improve the processing efficiency.3) To perform different functions, the BCI system works in three models:Interactive training mode to compete the optimization of the classifier model, offlineanalysis mode to finish the analysis and performance verification of signalprocessing algorithms, and the real-time control mode to realize the real-time controlbetween EEG and external device.4) Design a wireless smart car as the external device. They are transmittedwirelessly through the computer serial to control the smart car movement, When theEEG signals are classified and encoded. The results show that the designed BCIsystem can control external device effectively and the accuracy rate reached anaverage of80.42%. At last, on the condition that ensuring the accuracy of real-timecontrol, we add the shared control to prevent the car form hitting the barrier causedby the uncertainty of consciousness commands. |