Study On Tele-operated NAO Robot System Based On SSVEP

As a new type of human-computer interface, brain-computer interface (BCI) is an EEG-based communication and control system between human and computer or other electric devices, which could provide a communication and control channel for subjects with severe neuromuscular dysfunction to manage external events. This thesis has studied the data collection, extraction and identification of steady-state visual evoked potential (SSVEP), then designed and implemented a SSVEP-based BCI control system combined with tele-operated NAO robot control system. This thesis mainly involves two aspects about signal processing and system building.A data collection system for SSVEP has been built according to the studies of visual stimulator, and the data preprocessing method has been selected by comparing the butterworth band-pass filter and wavelet threshold denoising method. Three feature extraction methods, including singular value of the wavelet decomposition coefficients, subband energy of the wavelet packet ecomposition coefficients and colligate AR parameter model coefficient and empirical mode decomposition, have been proposed to process three kinds of VEP signals which collected from the data collection system. Three classifiers, including LVQ network classifier, BP neural network classifier based on Particle Swarm Optimization (PSO) and SVM classifier based on Genetic Algorithm (GA), have been utilised to process the eigenvector matrices. The results of the experiment show that the algorithms have achieved good classification accuracy.A SSVEP-based NAO robot tele-operation system has also been designed and built. Performance of the designed event-based robot control system has been improved by the robot feedback interaction interface and switching mode function. The system has been verified through a series of tasks as controlling the robot walking and placing objects.In conclusion, the signal extraction and the recognition algorithms are suitable for SSVEP-based BCI applications, and the system of using the EEG signal control for robot tele-operating is feasible, which also provides a new approach for clinical application.