Vehicle Trajectory Reconstruction Method for Urban Arterial Roads Based on Multi-Source Data Fusion

Area of Science:

• Transportation Science
• Data Fusion
• Traffic Engineering

Background:

• Vehicle trajectory data is crucial for urban traffic analysis and management.
• Challenges exist in obtaining comprehensive trajectory data due to low probe vehicle (PV) penetration and sensor coverage.
• Accurate trajectory data is essential for optimizing traffic flow and signal control.

Purpose of the Study:

• To develop a multi-source data fusion method for reconstructing vehicle trajectories.
• To address the limitations of low PV penetration rates and incomplete sensor data.
• To enhance the accuracy and smoothness of reconstructed vehicle trajectories for refined traffic management.

Main Methods:

• Trajectory state estimation for undetected vehicles categorized into four types.
• Reconstruction of initial trajectories using an extended Intelligent Driver Model.
• Particle filter-based self-optimization algorithm integrating fixed sensor data for iterative trajectory correction.

Main Results:

• The proposed method significantly reduces Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Root Mean Square Error (RMSE) compared to baseline methods.
• Performance is outstanding in complex traffic environments and under low PV penetration rates.
• Reconstruction errors decrease with increasing traffic density and PV penetration rates, with PV penetration showing a significant impact on accuracy.

Conclusions:

• The multi-source data fusion method demonstrates robustness and effectiveness in reconstructing vehicle trajectories.
• The approach provides critical technical support for refined urban traffic management and optimized decision-making.
• The findings highlight the importance of PV penetration rates in improving trajectory reconstruction accuracy.