Fault diagnosis for embedded system based on RBF neural network optimized by PSO

In the paper, RBF neural network optimized by particle swarm optimization algorithm is proposed to fault diagnosis for embedded system. The parameters of RBF neural network are selected by particle swarm optimization algorithm to solve the problem of the parameters selection of RBF neural network. In order to testify the superiority of RBF neural network optimized by optimized by particle swarm optimization algorithm in fault diagnosis for embedded system, traditional RBF neural network and BP neural network are applied to compare with the proposed RBF neural network optimized by particle swarm optimization algorithm. The comparison results show that the diagnosis accuracy of RBF neural network optimized by particle swarm optimization algorithm is the best methods among traditional RBF neural network, BP neural network and RBF neural network optimized by particle swarm optimization algorithm.