Research On Disturbance Rejection Of Spiking Neural Network Based On Synaptic Plasticity

With the rapid development of science and technology,the electromagnetic environment is becoming more and more complex.The shortcomings of the traditional anti-electromagnetic interference method are becoming increasingly prominent.Compared with the electronic system,the organism under the regulation of nervous system has the advantages of self-organization,self-adaption,reliability and disturbance rejection.Under the background of electromagnetic bionic protection,a feedforward spiking neural network based on synaptic plasticity was established,and the disturbance rejection of the feedforward spiking neural network under noise,electric field,magnetic field,and neuron damage was studied,which lays the theoretical foundation for improving the protection ability of the electronic system under the complex electromagnetic environment.The main work of this thesis is as follows:(1)The disturbance rejection of the neural network under noise,electric field and magnetic field was studied.Based on synaptic plasticity,a ten layer feedforward neural network was constructed.The firing rate and correlation between membrane potential of the output layer under these three kinds of disturbances were analyzed emphatically.The experimental results show that the noise,electric field and magnetic field have little influence on the output firing rate,and the certain noise,electric field amplitude and magnetic field amplitude have little effect on the neuron membrane potential of the output layer,the correlation between membrane potential is relatively large.It indicates that the spiking neural network based on synaptic plasticity has certain ability to resist noise,electric field and magnetic field interference.(2)The disturbance rejection of spiking neural network under neuron damage was studied.The firing rate and correlation between membrane potential of spiking neural network under different neuronal damage ratio were studied.The influence of neuron damage on the output layer was analyzed emphatically.The experimental results show that neuron damage has little influence on the output firing rate,and the certain neuron damage has little effect on the neuron membrane potential of the output layer,the correlation between membrane potential is relatively large.It indicates that the spiking neural network based on synaptic plasticity has certain ability to resist neuron damage interference.