BP神经网络在轮胎气压监测系统自学习匹配中的应用  

Application of BP Neural Network in Tire Pressure Monitoring System Self-Learning Matching

作　　者：王立强[1] 王斌[1,2] 王俊昌[3] 戴希[1] 韩宗奇 Wang Liqiang;Wang Bin;Wang Junchang;Dai Xi;Han Zongqi(Yanshan University, Qinhuangdao 066000;Suzhou Automotive Research Institute of Tsinghua University, Suzhou 215200;Anyang Institute of Technology, Anyang 455000)

机构地区：[1]燕山大学,秦皇岛066000 [2]清华大学苏州汽车研究院,苏州215200 [3]安阳工学院,安阳455000

出　　处：《汽车技术》2018年第5期20-24,共5页Automobile Technology

基　　金：国家重点研发计划(2016YFB0101102);燕山大学基础研究专项课题理工A类(14LGA019);江苏省青年基金项目(BK20160402)

摘　　要：提出了一种基于BP神经网络的轮胎气压监测系统轮胎换位自学习匹配方法。该方法基于间接式轮胎压力监测系统和轮胎受力特性对换位后的轮速信号特征进行分析,运用BP神经网络识别轮胎换位方式。通过采集轮胎换位后各车轮轮速数据对BP神经网络进行训练,从而实现神经网络对轮胎换位的准确识别,使得TPMS在无人工干预下可自行识别轮胎换位状态。道路试验结果表明,完成训练后的网络可实现对未换位、交叉换位、前后换位和循环换位的有效识别,准确率达97.52%。In this paper, a self-learning tire matching method based on BP neural network for TPMS was proposed. This method analyzed wheel speed signal feature based on tire force characteristic and indirect TPMS, and used BP neural network to identify the tire rotation. By acquiring wheels speed data of tire rotation, BP neutral network was trained, that enabled accurate recognition of tire rotation, helping the TPMS to automatically recognize tire rotation state without manual intervention. Road test results show that the trained BP neural network can effectively recognize the modes between no rotation, cross rotation, front-rear rotation and cycle rotation, with correct recognition rate up to 97.52%.

关 键 词：汽车轮胎 轮胎换位 胎压监测 BP神经网络 自学习 

分 类 号：U461.91[机械工程—车辆工程]