Research And Implementation Of Navigation Eeg Processing Based On Deep Learning

As a basic ability of the brain,navigation cognitive behavior is a mixture of various cognitive abilities,such as visual perception,direction perception,memory,and decision-making.Currently,little is known about the mechanism of its occurrence.Brain waves,as electrical signals triggered by neuronal activity in the brain,have become the main way to study brain activity due to their ability to characterize the state of brain activity and their high resolution and ease of collection.This project aims to study the navigation cognitive features of the brain,using the collected two EEG signals as the starting point,combined with signal processing techniques such as noise reduction filtering and deep learning techniques such as convolutional neural networks,to carry out a series of experiments.The specific content is as follows:Acquisition of navigation cognitive EEG data sets in this study.In order to study the characteristics of the human brain during navigation tasks,directional sensing EEG acquisition experiments and navigation planning EEG acquisition experiments were conducted.Then,two navigation EEG data sets were constructed using preprocessing algorithms such as independent component analysis and sliding window slicing methods.Propose a time domain composite attention neural network model in this study.Considering the high sampling rate of navigation signals,the model introduces multi time domain convolution to extract signal features under different time long Receptive field,then uses attention mechanism to allocate different channel weights,and finally extracts more advanced signal features through time convolution technology.After comparative experiments,the model achieved the best performance on the directional perception EEG dataset and the public motor imagery dataset.Based on the navigation planning task EEG dataset and the proposed classification model,the differences and similarities in EEG signals of 21 subjects during three different navigation tasks were compared and analyzed,and several conclusions on brain navigation cognition were obtained.The experimental process also provides a reference for studying navigation EEG characteristics.An EEG signal processing and feature recognition system was designed and developed.The system includes common functions such as EEG signal import and export,feature display,noise reduction processing,model training,classification and prediction,and has also been tested to verify the effectiveness of the system.