基于机器学习的间接式胎压监测算法研究

doi:10.3969/j.issn.1674-8484.2022.02.009

汽车安全与节能学报 ›› 2022, Vol. 13 ›› Issue (2): 290-299.DOI: 10.3969/j.issn.1674-8484.2022.02.009

基于机器学习的间接式胎压监测算法研究

于璐¹(), 唐亮¹^,^*(), 魏凌涛², 刘子俊²

1.北京林业大学工学院,北京100091,中国
2.清华大学车辆与运载学院,北京100084,中国

收稿日期:2022-01-11 修回日期:2022-03-17 出版日期:2022-06-30 发布日期:2022-07-01
通讯作者: 唐亮
作者简介:*唐亮(1981—), 女(汉), 湖北, 副教授。E-mail: happyliang@bjfu.edu.cn。
于璐(1994—),女(汉),河南,硕士研究生。E-mail： uluy_yulu@163.com。
基金资助:
湖北省重点实验室2021年开放课题资助项目(XDQCKF2021004)

Research on indirect tire pressure monitoring algorithm based on machine learning

YU Lu¹(), TANG Liang¹^,^*(), WEI Lingtao², LIU Zijun²

1. School of Engineering, Beijing Forestry University, Beijing 100091, China
2. School of Vehicle and Mobility, Tsinghua University, Beijing 100084, China

Received:2022-01-11 Revised:2022-03-17 Online:2022-06-30 Published:2022-07-01
Contact: TANG Liang

摘要/Abstract

摘要：

为了克服间接式胎压监测系统在复杂道路环境及可变行驶工况下辨识准确率低的问题,该文在仅依靠轮速传感器的基础上研究了基于机器学习的间接式胎压监测算法。对建立的轮胎刚性环模型进行分析,得到辨识方法的理论依据;使用递归最小二乘法(RLS)剔除轮速传感器误差,获得准确的轮速信号;提取时域和频域轮速信号特征,利用决策树剔除问题轮速信号,基于正常轮速信号的特征综合Bayesian分类器识别当前轮胎压力状态。结果表明：与直接使用Bayesian分类器相比,本研究提出的决策树和Bayesian分类器结合在一起的间接式胎压监测方法具有更高的准确率,其准确率可达96.36%。

关键词: 间接式胎压监测系统（ITPMS）, 轮胎振动模型, 最小二乘法(RLS), 决策树, Bayesian分类器

Abstract:

A novel tire pressure monitoring algorithm based on machine learning was presented to solve the problem that the current indirect tire pressure monitoring systems (ITPMS) achieved poor identification accuracy under complex road status and varying driving conditions, which was implemented only with the existing wheel speed sensors in vehicles. The theoretical basis of the indirect TPMS was introduced by analyzing the rigid tire model. And the inherent error generated in speed sensors was effectively removed by the recursive least square (RLS) method and accurate wheel speed signals was obtained. The features in time and frequency domains were extracted and a decision tree was used to eliminate the abnormal speed signals and a Bayesian classifier was used to identify the tire pressure conditions overall based on features of the normal speed signals. The result shows that this method in conjunction with decision trees indicates higher correctness during indirect tire pressure monitoring, where the accuracy of identification reaches 96.36% in comparison of the method only with a Bayesian classifier.

Key words: indirect tire pressure monitoring system (ITPMS), tire vibration model, recursive least square (RLS), decision tree, Bayesian classifier

中图分类号:

U467.1+4

于璐, 唐亮, 魏凌涛, 刘子俊. 基于机器学习的间接式胎压监测算法研究[J]. 汽车安全与节能学报, 2022, 13(2): 290-299.

YU Lu, TANG Liang, WEI Lingtao, LIU Zijun. Research on indirect tire pressure monitoring algorithm based on machine learning[J]. Journal of Automotive Safety and Energy, 2022, 13(2): 290-299.

图/表 14

图1 轮胎振动模型

轮胎纵向刚度,k_x	4 kN/m
轮胎旋转刚度,k_φ	15 (kN·m)/rad
胎面刚度,k_t	30 kN/m
悬架纵向刚度,k_s	20 kN/m
带束环质量,m_B	10.9 kg
轮辋质量,m_R	30 kg
带束环转动惯量,J_B	1.0 kg·m²
轮辋转动惯量,J_R	0.69 kg·m²
静态滚动半径,r_stat	0.35 m

表1 轮胎振动模型仿真参数

图2 理想状态下轮速信号的幅频特性曲线

图3 基于机器学习的间接式胎压监测系统框架

图4 时域轮速信号示意图

图5 轮速传感器误差示意图

图6 轮速传感器齿圈误差估计结果

图7 典型时域轮速信号

图8 典型时域轮速信号的频谱图

图9 原始频域轮速信号

图10 平滑频域轮速信号

图11 决策树训练结果

图12 Bayesian分类器算法流程图

图13 利用Bayesian分类器识别胎压力状态的结果

参考文献 18

[1]	孙春玲. 基于轮胎模型的轮胎压力监测系统研究[D]. 济南: 山东理工大学, 2007.
	SUN Chunling. Research on tire pressure monitoring system basing on the tire model[D]. Jinan: Shandong University of Technology, 2007. (in Chinese)
[2]	谭德荣, 王艳阳, 张莉. 基于卡尔曼滤波算法的间接胎压监测方法[J]. 农机化研究, 2007(12): 74-78.
	TAN Derong, WANG Yangyan, ZHANG Li. The research about the method of indirect tire pressure monitoring based on Kalman filter algorithms[J]. J Agri Mech Res, 2007(12): 74-78. (in Chinese)
[3]	王艳阳. 基于ABS轮速信号的间接胎压监测技术研究[D]. 济南: 山东理工大学, 2008.
	WANG Yanyang. Technology research on tire pressure monitoring based on anti-lock brake system information.[D]. Jinan: Shandong University of Technology, 2007. (in Chinese)
[4]	韩宗奇, 刘全有, 王立强, 等. 基于标准脉冲数比较法的汽车轮胎压力监测与报警系统研究[J]. 中国机械工程, 2010, 21(2): 240-244+251.
	HAN Zongqi, LIU Quanyou, WANG Liqiang, et al. Study on tire pressure monitoring and alarming system of automobile based on the comparison among standard pulse numbers[J]. Chin Mech Engi, 2010, 21(2): 240-244+251. (in Chinese)
[5]	苏静. 基于轮速特征分析的胎压估计算法研究[D]. 长春: 吉林大学, 2017.
	SU Jing. Research on tire pressure estimation algorithm based on wheel speed characteristic analysis[D]. Changchun: Jilin University, 2017. (in Chinese)
[6]	李敏. 基于车载OBD输出信息的间接式TPMS研究[D]. 秦皇岛: 燕山大学, 2015.
	LI Min. Study of indirect TPMS based on OBD information of vehicle[D]. Qinhuangdao: Yanshan University, 2015. (in Chinese)
[7]	钟响. 间接式轮胎压力监测系统研究[D]. 长春: 吉林大学, 2016.
	ZHONG Xiang. Research on indirect tire pressure monitoring system[D]. Changchun: Jilin University, 2016. (in Chinese)
[8]	谢秀芬. 基于频率估计实现间接式轮胎压力监测系统的关键技术研究[D]. 长沙: 国防科学技术大学, 2007.
	XIE Xiufen. Research on key technologies of indirect tire pressure monitoring system based on frequency estimation[D]. Changsha: National University of Defense Technology, 2007. (in Chinese)
[9]	单经纬. 基于频域特征分析的间接式胎压监测算法研究[D]. 长春: 吉林大学, 2016.
	SHAN Jingwei. The research of indirect tire pressure monitoring algorithm based on frequency domain feature analysis[D]. Changchun: Jilin University, 2016. (in Chinese)
[10]	WANG Biao, LEI Yaguo, LI Naipeng, et al. Deep separable convolutional network for remaining useful life prediction of machinery[J]. Mech Syst Sign Proc, 2019, 134: 106330.
[11]	郝鑫. 基于Bayesian估计的环境车辆感知[J]. 汽车实用技术, 2021, 46(20): 31-33+44.
	HAO Xin. Environmental vehicle perception based on bayesian estimation[J]. Autom Appl Tech, 2021, 46(20): 31-33+44. (in Chinese)
[12]	吴西涛, 魏超, 翟建坤, 等. 考虑横摆稳定性的无人车轨迹跟踪控制优化研究[J]. 机械工程学报, 2022, 58(6): 130-142. doi: 10.3901/JME.2022.06.130
	WU Xitao, WEI Chao, ZHAI Jiankun, et al. Study on the optimization of autonomous vehicle on path-following considering yaw stability[J]. J Mech Engi, 2022, 58(6): 130-142. (in Chinese)
[13]	查园园, 王亭岭, 上官伟. 基于Bayesian网络的列控车载设备故障诊断[J]. 北京交通大学学报, 2021, 45(5): 37-45.
	CHA Yuanyuan, WANG Tingling, SHANGGUAN Wei. Bayesian network-based fault diagnosis for on-board equipment of train control system[J]. J Beijing Jiaotong Univ, 2021, 45(5): 37-45. (in Chinese)
[14]	YANG Shaojie, LI Xiang, JIA Xiaodong, et al. Deep Learning-based intelligent defect detection of cutting wheels with industrial images in manufacturing[J]. Procedia Manufact, 2020, 48: 902-907.
[15]	ZHAO Rui, YAN Ruqiang, CHEN Zhenghua, et al. Deep learning and its applications to machine health monitoring[J]. Mech Syst Sign Proc, 2019, 115: 213-237.
[16]	Ludwig S A, Picek S, Jakobovic D. Classification of cancer data: Analyzing gene expression data using a fuzzy decision tree algorithm[M]// Operations Research Applications in Health Care Management.Springer, Cham, 2018: 327-347.
[17]	罗锲铭. 基于频率法的间接式轮胎压力监测系统研究[D]. 武汉: 武汉理工大学, 2015.
	LUO Qiming, The research of indirect tire pressure monitor system based on the frequency method[D]. Wuhan: Wuhan University of Technology, 2015. (in Chinese)
[18]	Suender R, Prokop G, Roscher T. Comparative analysis of tire evaluation methods for an indirect tire pressure monitoring system (ITPMS)[J]. SAE Int’l J Passenger Cars-Mech Syst, 2015, 8: 110-118.

基于机器学习的间接式胎压监测算法研究

Research on indirect tire pressure monitoring algorithm based on machine learning

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 14

参考文献 18

相关文章 4

编辑推荐

Metrics

本文评价

期刊信息

在线期刊

作者中心

审稿中心

联系我们

[1]	陈德海, 王昱朝, 孙仕儒, 雷志军. 基于无迹Kalman滤波算法的电池内部温度估计[J]. 汽车安全与节能学报, 2022, 13(1): 186-193.
[2]	李欢, 白中浩, 高文睿, 周启峰, 蒋彬辉. 基于决策树模型的电动自行车与SUV碰撞中骑车人头部响应[J]. 汽车安全与节能学报, 2021, 12(1): 43-51.
[3]	王超, 陈德海, 王昱朝, 朱正坤, 邹争明. 基于无迹 Kalman 滤波算法的动力电池荷电状态估计[J]. 汽车安全与节能学报, 2020, 11(3): 379-389.
[4]	王晓原，刘亚奇，张敬磊. 基于行程时间的人车特征动态辨识方法[J]. JASE, 2017, 08(01): 38-45.