Research On Cognitive Computing Based On Brain Computer Interface

Cognition refers to the process of acquiring knowledge,processing knowledge and applying knowledge,which is the basic psychological activity of human beings.The information input from the outside is transmitted to the brain through the human sensory organ.The brain processes all kinds of this information more deeply,turning it into a thought that is unique to the individual,and then redirecting human behavior through this activity to respond to stimuli.Cognitive science mainly studies the mechanism of biological nerves acting on information and knowledge in the process of perception,storage,analysis and expression,and then builds mathematical models on this basis.Because of the poor expansibility of intelligence that using the traditional exact logic operations such as fuzzy sets or rough sets to simulate the processing of uncertain information by human cognition.The cloud model is proposed,this model is used to describe the uncertain relationship between concept and extension.Brain-computer interface(BCI)technology is a communication control strategy which is independent of the peripheral nerves and muscles of the brain and is controlled directly by EEG signals.The aim of this paper is to extract the characteristics of human EEG signals in different behavioral states,and to obtain the behavioral state information represented by EEG signals through the analysis of the characteristics,so as to achieve the purpose of control.However,because of its high complexity,large amount of data and strong uncertainty,EEG analysis becomes more difficult.The research of cognitive computing method based on brain-computer interface aims to complete the information extraction and pattern recognition of EEG signal by using cognitive computing method.Therefore,this paper mainly includes the following aspects:(1)First of all,Noise reduction of EEG signals by several methods,compared and optimized design of various methods to find the best processing methods.And then selection of the channels through the relation between seat and specific activities,Finally,by feature extraction of the brain wave,we can find the signal features corresponding to all kinds of activities contained in the brain wave signal,so as to prepare for the next step.(2)The mathematical essence of cloud model is studied.The probability distribution function,reverse cloud algorithm and parameter selection of cloud model are further explored to find out the optimal algorithm model.It is used to test the reliability of the model and find out the advantages and disadvantages of different algorithms.Finally,the bidirectional cloud transform algorithm is used to simulate four kinds of human cognitive processes,and the characteristics of human cognition are analyzed through the simulation of this process.(3)According to the cognitive characteristics of existing EEG signals,combined with the transformation algorithm of bi-directional cloud cognition,taking naive Bayes classifier as an example,the pattern recognition of EEG signals can be realized by adding the processing ability of uncertainty to the data by the cognitive algorithm.Through the analysis and comparison of the results of the original method and the improved algorithm,the validity of the method is judged.