Masino Johannes, Foitzik Michael-Jan, Frey Michael, Gauterin Frank
Institute of Vehicle System Technology, Karlsruhe Institute of Technology (KIT), Kaiserstr. 12, 76131 Karlsruhe, Germany.
J Acoust Soc Am. 2017 Jun;141(6):4220. doi: 10.1121/1.4983757.
Tire road noise is the major contributor to traffic noise, which leads to general annoyance, speech interference, and sleep disturbances. Standardized methods to measure tire road noise are expensive, sophisticated to use, and they cannot be applied comprehensively. This paper presents a method to automatically classify different types of pavement and the wear condition to identify noisy road surfaces. The methods are based on spectra of time series data of the tire cavity sound, acquired under normal vehicle operation. The classifier, an artificial neural network, correctly predicts three pavement types, whereas there are few bidirectional mis-classifications for two pavements, which have similar physical characteristics. The performance measures of the classifier to predict a new or worn out condition are over 94.6%. One could create a digital map with the output of the presented method. On the basis of these digital maps, road segments with a strong impact on tire road noise could be automatically identified. Furthermore, the method can estimate the road macro-texture, which has an impact on the tire road friction especially on wet conditions. Overall, this digital map would have a great benefit for civil engineering departments, road infrastructure operators, and for advanced driver assistance systems.
轮胎路面噪声是交通噪声的主要来源，会导致人们普遍感到烦恼、干扰交谈以及影响睡眠。测量轮胎路面噪声的标准化方法成本高昂、使用复杂，且无法全面应用。本文提出了一种自动分类不同类型路面及其磨损状况以识别噪声路面的方法。该方法基于在正常车辆运行条件下获取的轮胎腔声音的时间序列数据频谱。分类器采用人工神经网络，能够正确预测三种路面类型，对于两种物理特性相似的路面，双向误分类情况较少。分类器预测新路面或磨损路面状况的性能指标超过94.6%。利用所提出方法的输出结果可以创建数字地图。基于这些数字地图，可以自动识别对轮胎路面噪声有强烈影响的路段。此外，该方法还可以估计道路宏观纹理，其对轮胎路面摩擦力有影响，尤其是在潮湿条件下。总体而言，这种数字地图将对土木工程部门、道路基础设施运营者以及先进驾驶辅助系统大有裨益。
