2022 Volume 8 Issue 4 Pages A_9-A_15
The Ministry of Land, Infrastructure and Transport of Japan has started a project, which is one of the policies to protect the citizens from traffic accidents, to utilize probe data such as ETC2.0 probe data in road projects. However, sudden braking data include both risky braking events caused by averting automobile accidents and non-risky braking events caused by deceleration behaviors at red light signals, 'STOP' signs, etc. Therefore, to lead an accurate traffic safety counter by using sudden brake data, researching the characteristics of sudden brake to identify true potential hazard spots are necessary. In this paper, to extract the real potentially dangerous spots (or areas), We focus on the wave form of driving recorder data. Basing on the analyzing results of the feature amounts of wave form by using multivariate analysis (Non-hierarchical cluster analysis and Decision Tree), we propose an index for determining risk events and non-risk events.