Road Rescue Demand Prediction for the Improvement of Traffic System Resilience

Abstract

Road rescue can provide rescue services for faulty vehicles, such as fuel delivery, tire replacement, battery connection, on-site repair, clearing, and towing, which plays an important role in reducing casualties and property losses in traffic accidents. Based on the historical data of road rescue, this paper analyzes the influencing factors of the road rescue demand and establishes a prediction model of the road rescue demand without data grouping. In order to further improve the prediction accuracy, the data are divided into nine groups according to the importance of the influencing factors, and nine submodels are established for the nine groups of data. When the influencing factors are known, the submodel corresponding to the most important influencing factor is selected to predict the road rescue demand. A case study in Beijing is used to verify the effectiveness and superiority of the proposed models, which can effectively predict the road rescue demand under various conditions, including the normal condition, the Spring Festival, National Day, the three-day holiday (e.g., Qingming, May Day, the Dragon Boat Festival, the Mid-Autumn Festival, and New Year’s Day,), and extreme weather (e.g., low temperature, high temperature, heavy snow, heavy rain, and rainstorm). The research findings can provide scientific basis for the rescue department to deploy rescue equipment and rescue personnel in advance, raise the efficiency and quality of rescue, and improve the resilience of the transportation system.