Early warning method for rear end traffic accidents in highway tunnel groups based on big data analysis

Aiming to minimize false alarms and decrease response times in rear-end accident detection, we introduce a big data analytics approach for highway tunnel clusters. Firstly, a deep learning framework based on uncertain association rules is used to mine big data on traffic flow in highway tunnels. Secondly, based on the three major characteristic parameters, a warning standard system for rear end collisions in highway tunnel groups is constructed. Finally, an improved artificial immune algorithm is adopted to develop a three-stage full process warning mechanism. By optimizing the detector matching strategy and introducing a dynamic mutation mechanism, accurate early warning of severe, moderate, and mild level accidents has been achieved. The experimental results show that the false alarm rate of the warning method proposed in this paper is less than 1% and the response delay is controlled within 0.5s.