基于图像特征和HMM的前方附着系数辨识  

Identification of adhesion coefficient ahead based on image feature and HMM

作　　者：袁朝春[1] 张海峰[1] 何友国[1] JIE Shen 陈龙[1] YUAN Chaochun;ZHANG Haifeng;HE Youguo;JIE Shen;CHEN Long(Automotive Engineering Research Institute,Jiangsu University,Zhenjiang,Jiangsu 212013,China;Department of Computer and Information Science,University of Michigan-Dearborn,Dearborn,Michigan MI 48128,USA)

机构地区：[1]江苏大学汽车工程研究院,江苏镇江212013 [2]密西根大学迪尔本分校计算机科学与信息科学系,密西根州迪尔本MI 48128

出　　处：《江苏大学学报（自然科学版）》2022年第4期373-380,393,共9页Journal of Jiangsu University：Natural Science Edition

基　　金：国家自然科学基金资助项目(51775247)。

摘　　要：为了辨识前方道路附着系数,利用灰度共生矩阵和HSV(hue,saturation,value)颜色空间提取了典型路面图像7种纹理特征参数,基于类比特性对Burckhardtμ⁃s模型进行改进,提出一种当前路面峰值附着系数实时估计算法,通过Carsim和Simulink联合仿真验证了其有效性和实时性.建立基于隐马尔可夫模型(hidden Markov model,HMM)的前方道路-轮胎附着特性模型,并验证所建模型的正确性.通过实车试验对HMM模型算法进行了验证.结果表明:HMM模型对前方道路-轮胎附着系数识别率达到90%以上,并且能够用于智能汽车自动紧急制动模块中,可以在前方恶劣路面环境下进行提前制动,有效缩短制动距离.To identify the road adhesion coefficient ahead,the grayscale co⁃occurrence matrix and HSV color space were used to extract the seven texture feature parameters of typical road images.Based on the similarity characteristics,the Burckhardtμ⁃s model was improved,and a real⁃time estimation algorithm of the current road peak adhesion coefficient was proposed.The effectiveness and real⁃time performance were verified by Carsim and Simulink co⁃simulation.A road⁃tire adhesion characteristic model was established based on hidden Markov model(HMM),and the correctness of the model was verified.The HMM model algorithm was verified through real vehicle experiments.The results show that the recognition rate of HMM model is more than 90%for the road⁃tire adhesion coefficient ahead,and the proposed model can be used in the automatic emergency braking module of smart cars.The vehicle can brake in advance under the front harsh road environment,which effectively shortens the braking distance.

关 键 词：道路-轮胎附着特性 路面图像特征 μ⁃s曲线模型 隐马尔可夫模型 预测算法 

分 类 号：U463.5[机械工程—车辆工程]