Trajectory prediction system and method, electronic equipment and readable storage medium

Abstract

The invention provides a trajectory prediction system and method, electronic equipment and a readable storage medium, and the method comprises the steps: obtaining the historical position of an automatic vehicle in a world coordinate system, the historical relative position of a to-be-predicted object relative to the automatic vehicle, and the environment information of the to-be-predicted object, based on the historical position of the automatic vehicle in the world coordinate system, the historical relative position of the to-be-predicted object relative to the automatic vehicle and the environment information of the to-be-predicted object, obtaining the historical positions of the to-be-predicted object and surrounding obstacles in the world coordinate system; according to the historical positions of the to-be-predicted object and the surrounding obstacles in the world coordinate system, obtaining and semantically extracting the relative historical trajectory of the to-be-predicted object based on the historical relative positions of the to-be-predicted object relative to the surrounding obstacles, and combining the relative historical trajectory with the lane information subjected to semantic extraction to obtain a semantic historical trajectory; and inputting the semantic historical trajectory into the trajectory prediction model to perform trajectory prediction of the to-be-predicted object. According to the method, the unmanned driving computing power is greatly improved, the computing consumption in the unmanned driving operation is minimized, the prediction precision can be ensured, and the engineering requirements are better met.

Description

technical field [0001] The invention relates to the field of automatic driving, in particular to a trajectory prediction system and method, electronic equipment and a readable storage medium. Background technique [0002] With the development of the field of intelligent transportation, the prediction algorithm of the trajectory of moving objects is of great significance in the field of path planning. By predicting the trajectory of the moving object, path planning can be performed when the possible future trajectory of the moving object is known, which is beneficial to prevent accidents such as collisions. [0003] In the existing technology, the position of surrounding vehicles can be obtained through perception, and the trajectory prediction of the object to be predicted can be performed according to the vehicle dynamic model, Kalman filter (KF), Gaussian mixture model (GMM), hidden Markov model (HMM), etc., Such methods can establish accurate track measurement models or ...

AI Technical Summary

Problems solved by technology

[0003] In the existing technology, the position of surrounding vehicles can be obtained through perception, and the trajectory prediction of the object to be predicted can be performed according to the vehicle dynamic model, Kalman filter (KF), Gaussian mixture model (GMM), hidden Markov model (HMM), etc., Such methods can establish accurate track measurement models or vehicle state transition models, but they can only predict the state of the next movement of objects, and cannot meet the requirements of automatic driving for long-term prediction of future obstacles

In addition, by observing the trajectory of obstacles for a period of time, the long-short-term memory network (LSTM) can be used to predict the trajectory of the object to be predicted. This method has a large cumulative error. Although it can achieve a long-term prediction, it only relies Due to the location of obstacles, the predicted future trajectory often loses the influence of environmental changes on it.

Images