Real-time vehicle trajectory prediction method for signalized intersection

Abstract

The invention provides a real-time vehicle trajectory prediction method for a signalized intersection. The method comprises the following steps that trajectory data during phase switching of the signalized intersection is acquired, and historical trajectory libraries are built according to the vehicle stopping and passing decisions and the following vehicle state respectively; the stopping and passing decisions of vehicles are pre-judged on the basis of statistical data of historical trajectory characteristic indexes; the single vehicles and the following vehicles are pre-estimated on the basis of the historical trajectory libraries by means of a K-NN model and an IDM following vehicle model respectively; and trajectory prediction is conducted by means of Kalman filtering. According to thetrajectory prediction method, multiple steps sizes of trajectory prediction can be conducted on the vehicles at the signalized intersection, the dangerous driving behaviors of the vehicles can be recognized in advance, the method can be applied to a real-time signal control strategy or vehicle-mounted early warning system, and the potential accident risk caused by the dangerous driving behaviorsis eliminated actively.

Description

technical field [0001] The invention belongs to the field of traffic information, in particular to a real-time vehicle track prediction method suitable for signal-controlled intersections. Background technique [0002] Accurate prediction of vehicle trajectory is to prepare for the early identification of dangerous driving behavior. If the vehicle movement at the next moment can be accurately predicted in advance, the recognition effect of dangerous driving behavior will be greatly improved, and the driver will be reminded in time to take effective collision prevention measures in time. [0003] At present, there are not many studies on the use of trajectory data. The reason is that the acquisition of trajectory data is difficult, the cost is high, and the technology brought about by the multi-dimensional and dynamic driving characteristics (speed and acceleration of changing speed and acceleration) of various types of vehicles. Complexity. Research in the field of behavio...

AI Technical Summary

Problems solved by technology

However, this type of prediction algorithm can only perform single-step prediction on trajectory data, and is mostly suitable for arterial roads or expressways, and is not very applicable to signal-controlled intersections with more complex situations.

Images