摘要
为实现路面抗滑性能非接触式测量,提出应用新型区域三维纹理特征来表征沥青路面形貌构造并评估路面抗滑性能.使用便携式高分辨率三维激光扫描仪采集不同类型沥青路面纹理数据,同时使用动态摩擦系数测试仪采集其路面抗滑性能数据,并分别使用70和15 km/h时的动态摩擦系数值代表高速与低速状态下的路面抗滑性能.通过相关性分析和多元线性回归,发现路面抗滑性能与多个区域三维纹理特征参数的共同作用有关.建立前馈神经网络预测模型,使用多个区域三维纹理特征参数预测高速与低速状态下的路面抗滑性能.结果表明,区域三维纹理特征参数对动态摩擦系数测试仪在70 km/h时测得的路面抗滑性能预测能力为77%,对15 km/h时测得的路面抗滑性能预测能力为69%,证实了区域三维纹理特征参数与路面抗滑性能之间存在非线性联系.
To realize contactless skid resistance measurement,novel three-dimensional(3D)areal texture parameters are proposed to characterize the pavement surface topography and evaluate the skid resistance.A portable ultrahigh-resolution 3D laser scanner is used to collect various asphalt pavement texture data.A dynamic friction tester(DFT)is used to collect the asphalt pavement skid resistance data in parallel.The friction coefficient collected by DFT at speeds of 70 and 15 km/h represents the pavement skid resistance at the higher speed and lower speed,respectively.The joint action of the 3D areal texture parameters with each other on skid resistance is found out via the correlation analysis and multiple linear regression.The skid resistance at both higher and lower operating speeds is characterized by the 3D areal texture parameters using a multilayer feed-forward neural network model.Results indicate that the 3D areal texture parameters account for 77%contributions to higher speed skid resistance and 69%contributions to lower speed skid resistance.It proves that there is a non-linear relationship between the 3D areal texture parameters and skid resistance.
作者
彭毅
李强
战友
杨广伟
王郴平
Peng Yi;Li Qiang(Joshua);Zhan You(Jason);Yang Guangwei;Wang Kelvin C.P.(School of Civil Engineering,Southwest Jiaotong University,Chengdu 610031,China;Key Laboratory of Highway Engineering of Sichuan Province,Southwest Jiaotong University,Chengdu 610031,China;School of Civil and Environmental Engineering,Oklahoma State University,Stillwater,OK 74078-5013,USA;Guangdong Provincial Academy of Building Research Group Co.,Ltd.,Guangzhou 510500,China)
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2020年第4期667-676,共10页
Journal of Southeast University：Natural Science Edition
基金
国家自然科学基金资助项目(U1534203,51478398)
中国博士后科学基金资助项目(2019M663557)
国家留学基金委资助项目(201607000101)。
关键词
路面纹理
抗滑
区域纹理参数
神经网络
pavement texture
skid resistance
areal texture parameter
neural network
