Inference Method for Traffic Accident in Highway Monitoring Blind Based on Running Speed

Abstract

This study proposes and validates an inference method to detect abnormal traffic events in highway monitoring blind spots by analyzing vehicle speed data from adjacent monitoring areas. Utilizing simulation tools such as Vissim and LumenRT, various realistic traffic scenarios were designed to collect comprehensive data on vehicle speeds. Cross-correlation and Pearson correlation coefficients were employed to identify time lags and quantify the relationship between speeds in different sections. The results indicated a significant correlation under normal conditions and a marked decrease during abnormal events. This method provides a reliable approach to enhancing the detection of traffic anomalies, improving highway safety and efficiency. The implications of this research suggest that integrating such methods into existing traffic management systems can offer a more comprehensive surveillance framework, with potential for future enhancements through real-world validation and the inclusion of additional variables.