Chapter 11: Road Safety Detection Modeling Based on Vehicle Monitoring Data in China

By investigating information of an average speed of vehicles and characteristics of speed distribution and speed volatility, this chapter develops a road safety detection system model in relating to real-time speed. This model analyzes the risk degree across double dimensions of roads and vehicles. The dimension of roads focuses on characters of speed dispersion and distribution, constructs coefficients of variation for different roads, and recognizes multimodal distributions through nuclear density. The dimension of vehicles focuses on characteristics of speed changing in moving succession and analyzes outliers of speed, speed clusters, and speed variance in adjacent roads. The calculating outcomes of both above dimensions are combined to be “risk factors” for road safety, which can judge risk degree of both road and vehicle, as well as issue real-time warning for road and vehicle. The empirical results show that this model provides a strong theoretical support for road safety detection based on vehicle monitoring data.