Abstract: This paper presents a cooperative spectrum sensing algorithm based on neural network with wolf pack optimization, and the neural network in this algorithm has the optimal structure. Based on the cooperative spectrum sensing algorithm with self-organizing neural network, this algorithm elaborates the generation of training samples and trained neural network. It also uses the wolf pack optimization method to optimize weight matrix after the training of neural network. The simulation results show that spectrum sensing algorithms based on self-organizing neural network with wolf pack optimization has better performance than the original algorithm.

Key words: neural network, spectrum sensing, cooperative, wolf pack algorithm

摘要： 提出了一种基于狼群优化的人工神经网络频谱感知方法，实现了具有神经网络最优结构的神经网络频谱感知算法。该算法在包含自组织神经网络的频谱感知算法的基础上，具体阐述了训练样本的生成，神经网络的训练以及对神经网络训练阶段结束后所得到的权值矩阵运用狼群优化方法进行进一步的优化处理的过程。实验结果表明，狼群优化的自组织神经网络频谱感知算法与自组织神经网络的频谱感知算法相比，具有更好的频谱感知性能。

关键词: 神经网络, 频谱感知, 协作式, 狼群算法