Vehicle Trajectory Prediction Method Based on License Plate Information Obtained from Video-Imaging Detectors in...

Area of Science:

• Transportation Engineering
• Data Science
• Urban Planning

Background:

• Vehicle license plate data from video detectors offers insights into traffic patterns and driver behavior.
• Accurate vehicle trajectory prediction is crucial for efficient urban road network management.

Purpose of the Study:

• To propose a real-time vehicle trajectory prediction method using historical license plate data.
• To enhance trip chain continuity and accuracy in urban environments.
• To develop a predictive model for vehicle movements at intersections.

Main Methods:

• Acquisition of vehicle trip chains using road network topology and intersection data.
• Implementation of the Dijkstra algorithm for trip chain compensation with missing data.
• Calculation of turning state transition matrices from historical trip data.
• Development of a k-step prediction model for downstream intersection possibilities.

Main Results:

• The proposed method achieves an average accuracy of 0.72 for one-step trajectory prediction with only 200 historical training samples.
• Demonstrated effectiveness in a realistic urban road traffic scenario.
• Significant performance improvements in vehicle trajectory prediction.

Conclusions:

• The developed method provides a robust approach for real-time vehicle trajectory prediction.
• The technique effectively handles data gaps and improves the integrity of vehicle trip chains.
• This contributes to better understanding and management of urban traffic flow.