Vehicle track prediction method considering continuous interaction information

Abstract

The invention discloses a vehicle trajectory prediction method considering continuous interaction information, and the method comprises the steps: fully extracting vehicle motion information and inter-vehicle space interaction information in a vehicle trajectory data set through a built network model, and precisely predicting the future trajectory of a vehicle; firstly, a vehicle track data set is processed; then, designing a trajectory prediction model of the vehicle; and finally, constructing a loss function, and training a trajectory prediction model of the vehicle. According to the method, when space interaction between vehicles is modeled, space interaction information between the vehicles at each historical moment is extracted, time relevance of the space interaction information between the vehicles is considered, a time relevance extractor is designed, the time relevance of the space interaction information between the vehicles is fully captured, and the time relevance of the space interaction information between the vehicles is extracted. Constraint is added to the predicted trajectory of the vehicle, so that the vehicle trajectory predicted by the network is more accurate.

Description

technical field [0001] The invention belongs to the technical field of intelligent driving, and relates to a vehicle trajectory prediction method, more specifically, to a vehicle trajectory prediction method considering continuous interactive information. Background technique [0002] With the development of intelligent driving technology, driving safety has received extensive attention from scholars at home and abroad. In order to drive safely and efficiently in the increasingly complex road traffic environment, smart cars not only need to accurately locate the surrounding vehicles in real time, but also should accurately understand and predict the driving behavior and future trajectories of surrounding vehicles, which will affect the decision-making of smart cars. and exercise risk assessment is very important. However, the driving trajectory of the vehicle is not only affected by the historical trajectory of the own vehicle, but also the interaction between the vehicles ...

AI Technical Summary

Problems solved by technology

However, the driving trajectory of the vehicle is not only affected by the historical trajectory of the own vehicle, but also the interaction between vehicles cannot be ignored, which greatly increases the difficulty for smart cars to predict the future trajectory of surrounding vehicles.

[0003] At this stage, more and more vehicle trajectory prediction methods take into account the interaction between vehicles. The advantage is that the spatial interaction information between vehicles is used as the constraint condition of the vehicle movement to be predicted, which limits the scope of vehicle trajectory prediction. However, the existing How to explicitly model the spatial interaction between vehicles is a challenge for predicting vehicle trajectories.

Images