不同拓扑结构脉冲神经网络抗扰功能对比分析

在复杂多变的电磁环境下,电子系统的传统抗电磁干扰方式的不足日益凸显。借鉴生物体在自适应抗扰方面的优势,寻求新的思路对提高电子系统的可靠性具有重要意义。以小世界网络和随机网络为拓扑结构,构建以Izhikevich神经元为节点,兴奋性和抑制性突触可塑性共同调节的脉冲神经网络。以脉冲神经网络的放电率和膜电位相关性为抗扰指标评估小世界脉冲神经网络和随机脉冲神经网络的抗扰功能,并将两种网络的抗扰功能进行对比。实验结果表明:在高斯白噪声刺激下,小世界脉冲神经网络具有一定的抗扰功能和抗扰范围,其抗扰功能优于随机脉冲神经网络。

COMPARATIVE ANALYSIS OF ANTI-INTERFERENCE FUNCTION OF IMPULSIVE NEURAL NETWORKS WITH DIFFERENT TOPOLOGIES

In the complex and variable electromagnetic environment,the shortage of traditional anti-electromagnetic interference methods of electronic systems have become increasingly prominent.Learning from the advantages of living organisms in adaptive anti-interference,it is of great significance to seek new ideas to improve the stability of electronic systems.The small world network and random network were used as the topological structure,and the Izhikevich neurons were used as nodes,and the impulse neural network was adjusted together with excitatory synaptic plasticity and inhibitory synaptic plasticity.The anti-interference function of small world pulse neural network and random impulse neural network was evaluated by the firing rate and membrane potential correlation of pulsed neural network as the anti-interference index,and the anti-interference functions of the two networks were compared.The experimental results show that under the stimulation of Gaussian white noise,the small world pulse neural network has certain anti-interference function and anti-interference range,and its anti-interference function is better than random impulse neural network.