A kind of obstacle trajectory prediction method and device

Abstract

This specification discloses a trajectory prediction method and device for obstacles. The embodiment of this specification determines the global interaction characteristics of the vehicle and each obstacle under the joint action according to the historical state information and current state information of the vehicle itself, each obstacle, and the future trajectory planned by the vehicle itself; and according to the global interaction characteristics and the pending Predict the current state information of the obstacle, determine the individual interaction characteristics of the obstacle to be predicted; predict the future trajectory of the obstacle to be predicted through the individual interaction characteristics and the environment information of the vehicle itself. The embodiment of this specification uses the global interaction feature to characterize the future interaction results between the vehicle and each obstacle. When predicting the trajectory of the obstacle to be predicted, not only the interaction between the vehicle and each obstacle is considered. The impact of the trajectory also refers to the future trajectory planned by the vehicle itself, making the predicted future trajectory closer to the actual trajectory.

Description

technical field [0001] This specification relates to the field of unmanned driving technology, and in particular to a method and device for trajectory prediction of obstacles. Background technique [0002] At present, as an important part of artificial intelligence technology, the intelligentization of vehicles plays an increasingly prominent role in social production and life, and has become one of the main directions to guide the development of transportation technology. [0003] When planning the path of unmanned vehicles and vehicles with assisted driving functions (hereinafter collectively referred to as "vehicles"), in order to make the vehicle drive safely, it is necessary for the vehicle to avoid surrounding obstacles. Obstacles include static obstacles and dynamic obstacles. Since the static obstacles are stationary, it is easy for the vehicle to avoid the static obstacles. However, if the vehicle is to avoid dynamic obstacles accurately, it is necessary to predict...

AI Technical Summary

Problems solved by technology

Due to the complexity of the actual traffic conditions, there may be interactions between multiple dynamic obstacles and between dynamic obstacles and vehicles (for example, the motion state of a dynamic obstacle will be affected by another dynamic obstacle that is closer to it. or the impact of the vehicle's motion state, etc.), therefore, the accuracy of predicting the future trajectory of dynamic obstacles through the above method is poor, resulting in a lack of rationality in the results of vehicle path planning

Images