Data mining for safety transportation by means of using Internet survey Internet survey may be one of the effective means to collect big data from the real world. Collected data may realize meaningful analysis of targeted field. Intelligent Transportation (hereinafter: ITS) is one of smart city applications which bring us safety driving as well as comfortable driving by mitigation of the traffic congestion. This study proposes an example of vehicle-infrastructure cooperative function which would be incorporate into vehicle safety system for smart city application. Driver’s state adaptive driving support safety function may be one of key functions that can bring road traffic safety in combination with road infrastructure interactively. Consequently this study clarified root cause of traffic accidents by analysing data of experiences on traffic incidents. Data was collected by two methods, one was collected through direct interview, and the other was collected through Internet. From the analysis, haste, distraction were major factor of a driver’s psychosomatic states just before traffic incidents. As an alternative characteristic of driver’s cognitive distraction, this study acquired physiological information, which were movements of eyes and head, and heart rate. A method of detecting driver’s cognitive distraction was established by using pattern recognition, which were AdaBoost and Error-Correcting Output Coding (ECOC). This study proposes driver’s psychosomatic adaptive driving support safety function in combination with ITS service for smart city applications.