改善pso 算法结合flann 在传感器动态建模中的应用.pdfVIP

改进 PSO 算法结合 FLANN 在传感器动态建模中的应用 1,2 3，4 2 张媛媛 ，徐科军 ，许耀华 (1，合肥工业大学仪器科学与光电工程学院，安徽省合肥市，230009 2，安徽大学电子科学与技术学院，安徽省合肥市，230039 3，合肥工业大学电气与自动化工程学院，安徽省合肥市，230009 4，工业自动化安徽省工程技术研究中心，安徽省合肥市，230009） 摘要：将改进的粒子群优化（PSO）算法和函数联接型神经网络(FLANN)相结合，实现传感器的动态线性建 模。利用传感器的动态标定实验数据，首先训练 FLANN 神经网络，网络训练结束后的权值作为粒子群中某 个粒子的初始值，而后利用改进的 PSO 算法继续寻优，得到的全局最优值即为所求的传感器动态模型的系 数。实验结果表明，该方法结合了PSO和FLANN两者的优点，建模精度高。 关键词：MAF传感器；粒子群优化算法；函数联接型神经网络；建模 中图分类号：TP212.6 文献标识码：A Dynamic Modeling Approach Based on Improved PSO and FLANN for Sensor 1，2 3，4 2 Zhang YuanYuan ,Xu KeJun ,Xu Yaohua (1, School of Instrument Science and Photoelectric Engineering, Hefei University of Technology, Hefei 230009 2,School of Electronic Science and Technology, Anhui University, Hefei 230039 3，School of Electrical and Automation Engineering ,Hefei University of Technology ,Hefei 230009 4 ，Engineering Technology Research Center of Industrial Automation, Anhui Province, Hefei 230009) Abstract: In this paper, the dynamic linear model of sensor is founded based on the improved particle swarm optimization (PSO) algorithm and function link artificial neural network (FLANN). According to measurement data in dynamic calibration ,the weights of network is used to initialize one particle station of whole particel swarm when the training of FLANN has finished.Then the improved PSO algorithm is applied ,the global best particle station of particle swarm is the coefficient of sensor’s dynamic model what we need. The experiment results show that this approach has excellence of both PSO and FLANN, meanwhile it has better precision.