A Sensor-Data-Driven Proactive Accident Detection and Traffic Prediction Method Based on Lane-Level Grid Partitioning and a Three-Dimensional Markov Model
The proposed sensor-driven method achieves lane-level accident detection and traffic prediction with high accuracy by fusing historical and real-time data within a three-dimensional Markov model and proves effective in diagnosing other traffic anomalies, such as congestion, and for continuous monitoring of roadway inci...