Understanding Scenarios for Cooperative V2V Active Safety Applications Using Connected Vehicle Datasets

With the increasing experimental deployment of cooperative vehicular communication systems in the real world, large amounts of naturalistic driving data of connected vehicles are collected. These data can provide us with potential insights into more realistic scenarios for cooperative active safety V2V applications than the conceptual scenarios before the deployment. In this paper, a data analytic methodology is proposed to extract critical information related to scenarios for cooperative V2V active safety applications from data available of the safety pilot model deployment (SPMD). Lateral accelerations are investigated and extreme events are identified by driving volatility metrics. Then, the context of those extreme events, such as on-board radar readings, are extracted and analyzed by K-means algorithm. The clustered results and corresponding scenarios for cooperative V2V active safety applications are further discussed. These results could benefit design and testing for cooperative V2V active safety applications.