Collision Avoidance Warning Algorithm Based on Spatiotemporal Position Prediction of Vehicles at Intersections

Aiming at the high false alarm rate of vehicle collision avoidance algorithms at intersections controlled by traffic lights, a vehicle collision avoidance warning algorithm based on vehicle spatiotemporal position prediction (SPPWA) is proposed. The algorithm first obtains real-time data information such as the heading angle and global positioning system (GPS) coordinates of the two vehicles from the OnBoard Unit (OBU), and then the data is preprocessed by different filtering methods, and then excludes the data information that the two vehicles cannot collide. Finally, the filtered data is used to predict the spatiotemporal position of the vehicle before the two vehicles reach the collision point and determine whether the vehicle will collide. The algorithm is verified in three vehicle crash scenarios through PreScan and Matlab/Simulink co-simulation. The experimental results show that after the data are preprocessed by Kalman filtering, the algorithm has the lowest false alarm rate in the three scenarios. It can effectively improve the driving safety of vehicles at intersections.