基于路面构造特性的附着系数识别研究  

Research on Identification of Adhesion Coefficient by Pavement Structural Characteristics

作　　者：彭鹏峰[1] PENG Pengfeng(Guangdong Polytechnic of Industry and Commerce,Guangzhou 510510)

机构地区：[1]广东工贸职业技术学院,广州510510

出　　处：《广东公路交通》2022年第2期7-11,22,共6页Guangdong Highway Communications

基　　金：广东工贸职业技术学院科技项目(编号:2020-ZK-11)。

摘　　要：为提高汽车主动防碰撞系统的预警精度,提出了一种基于路面构造特性的附着系数识别方法。首先通过激光扫描仪扫描不同类型和不同磨耗程度的沥青路面,获取路表纹理形貌的三维坐标,在Matlab中编制程序计算出MTD、MPD等7个路面构造特征参数数值;然后对特征参数进行相关性分析,选取MTD、MPD、Rq和Δq作为评价路面附着系数的代表性表征指标;最后应用BP神经网络建立附着系数与代表性表征指标之间的关系模型,并用实测数据对模型进行训练和验证。结果表明:通过MTD、MPD、Rq、Δq等四个路面构造特征参数及建立的神经网络模型,能够较好地预测路面附着系数。In order to improve the pre-warn precision of vehicle active anti-collision system, in this paper, a adhesion coefficient recognition method based on pavement structural characteristics has been proposed. Firstly, the three-dimensional coordinates of road surface texture have been obtained by scanning different types of asphalt pavement with different wear degree by laser scanner. Programming in Matlab has been adopted to calculate the MTD, MPD and other seven pavement structure index parameters.Then, through correlation analysis of texture indexes, MTD, MPD, Rq and Δq have been selected as the representative indexes of evaluating the adhesion coefficient of pavement;Finally, the relationship model between the adhesion coefficient and the texture feature indexes with BP neural network has been established, and the model has been trained and verified with the measured data. The results have shown that: through MTD, MPD,Rq and Δq these four road table texture feature parameters and the neural network model can identify the road adhesion coefficient well.

关 键 词：附着系数 路面构造 BP神经网络 LM算法 

分 类 号：U416.2[交通运输工程—道路与铁道工程]