New nonlinear identification method of platinum resistance sensor based on IPSO-RBFNN

Abstract

A new nonlinear identification method of the platinum resistance sensor based on radial basis function neural network using a improved particle swarm optimization algorithm is proposed to settle its nonlinear problem. The particle swarm optimization algorithm is improved by introducing the shrinkage factor and the particle variation factor. The function of the particle fitness is achieved based on the distance between the actual neural network output values and the expected output values. Decode the global optimum value in the swarm searching space as the initial value of network parameters. The simulation shows that the new nonlinear identification has better nonlinear identification accuracy and stability. It is proved that the method is effective and feasible.