A Data-Driven Approach for Driving Safety Risk Prediction Using Driver Behavior and Roadway Information Data

Future roadways will have a mix of autonomous and automated vehicles with regular vehicles that require human operators. To ensure the safety of all the road users in such a network, it is necessary to enhance the performance of the present Advanced Driver Assistance System (ADAS) for lower classes of vehicles. Real-time driving safety risk prediction is an essential element of an ADAS. In this paper, we propose a novel data-driven approach to predict traffic safety risk that can be customized for individual drivers by including driver-specific variables. In particular, we have used the elastic net regularized multinomial logistic regression and data from the Second Strategic Highway Research Program (SHRP 2) naturalistic driving study to build the predictive models. This paper rigorously examines the variables in the data set and performs data preparation and feature engineering steps to enhance the prediction performance with respect to model predictors. The model produces good results, and model adaptation/extensions for further improvements are discussed in the conclusion section. Two versions of the model are presented according to the level of warnings that the model can generate based on driving conditions. This paper provides a brief overview of the potential applications of the work.