Real-time wet/dry road surface discrimination using rolling noise acoustic footprint analysis

Road surface changes due to the weather conditions have a major influence in driving safety. To avoid accidents, drivers must adapt their driving style according to the road status. This adaptation mechanism depends on the driver ability to detect the pavement conditions. Unfortunately, this is often a difficult task. A system capable of notifying the driver about changes on the road surface due to weather aspects, could improve this adaptation mechanism, thus reducing accident risk. An electronic system capable of estimating the road surface status between dry and wet status is presented. The system is mainly based on the analysis of the tyre/road noise generated during driving. The sound emission pattern of the tyre/road interaction changes depending on whether the pavement is dry or wet. Thus discriminating the tyre/road noise acoustic footprint, it is possible to estimate the road status. Proposed system captures and digitizes tyre/road noise. Rolling noise is pre-processed extracting its spectral components to construct the feature vectors. Feature vectors are processed by a Support Vector Machine (SVM) based classifier. Then the classifier outputs the road surface status estimation. The first implementation of the system, using Matlab along with pre-recorded tyre/road noise, shows high success rates (around 91 %). These promising results are leading to the development of a hardware prototype, tightly integrated with the vehicle, and capable of computing the classification algorithms in real time.