汽车制动系统可靠性预测的模糊神经网络研究  

Research on the Fuzzy Neural Network used for Reliability Prediction of Automotive Braking System

作　　者：倪泉[1] 葛友华[2] 王斌[2] 

机构地区：[1]常州大学机械工程学院,江苏常州213164 [2]盐城工学院机械工程学院,江苏盐城224001

出　　处：《自动化仪表》2016年第1期34-37,共4页Process Automation Instrumentation

基　　金：江苏省自然科学基金面上资助项目(编号:BK20131221)

摘　　要：计时传统模糊神经网络算法在汽车制动系统(ABS)可靠性预测中存在预测精度不高、误差较大等问题,提出了一种基于优化隶属函数的改进模糊神经网络算法。采用偏移优化方法对模糊控制算法的隶属函数进行改进;引入粒子群算法进行自适应惯性权重的寻优能力、收缩因子的收敛速度优化;最后与模糊神经网络算法融合,调整原算法的中心值、宽度值和连接权值,避免原算法在汽车制动系统可靠性预测中陷入局部最小值。仿真实验表明,改进的模糊神经网络算法具有比传统神经网络算法和模糊控制算法更小的预测误差。To overcome the disadvantages of traditional fuzzy neural network algorithm in reliability prediction of automotive braking system, e. g. ,low prediction accuracy and large error, etc. , the improved fuzzy neural network algorithm by optimizing membership function is proposed. By adopting the method of offset optimization, the membership function of fuzzy control algorithm is improved, and the particle swarm algorithm is introduced to optimize adaptive inertia weight optimization capability and convergence rate of contraction factor,finally the fuzzy neural network algorithm is merged to adjust the central value, width value, and connection weight value of original algorithm, to avoid falling into local minimum in reliability prediction of automotive braking system with original algorithm. The simulation shows that the improved fuzzy neural network algorithm offers smaller prediction error than traditional neural network algorithm.

关 键 词：汽车制动系统(ABS) 可靠性预测模糊神经 网络粒子群算法(PSO) 函数优化 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术] TH6[自动化与计算机技术—计算机科学与技术]