Motion track prediction method and device, computer equipment and storage medium

Abstract

The invention relates to a motion track prediction method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring a current motion track of a traffic participant and a current lane map related to the current track; according to the number of the track points on the current motion track, the first position information of each track point, the speed information and the identifier of the corresponding traffic participant, constructing the characteristics of the current traffic participant; constructing features of the current lane map according to the number of lane points of each lane map in the current lane map, the second position information and direction information of each lane point and the identifier of the corresponding lane; and performing high-dimensional embedded feature learning according to the features of the current traffic participant and the features of the current lane map to obtain target feature data considering the correlation between track points, between track points and lane points, and between lane points, and performing track prediction according to the target feature data to obtain a predicted track within a specified duration. By adopting the method, the accuracy of trajectory prediction can be improved.

Description

technical field [0001] The present application relates to the technical field of automatic driving, and in particular to a motion trajectory prediction method, device, computer equipment and storage medium. Background technique [0002] With the development of computer technology, automatic driving technology has emerged, which is gradually applied in various fields, such as ore mining, construction site inspection, substation inspection, etc.; in the process of automatic driving, it is necessary to predict the driving of other traffic participants The trajectory is used to plan the current driving trajectory to ensure that the autonomous driving can reach the destination safely and accurately. At present, Transformer technology can be applied to trajectory prediction tasks; Transformer is a deep network model that does not use recurrent neural network (RNN) or convolutional network (CNN), and fully represents input and output data through self-attention (self-attention) mec...

AI Technical Summary

Problems solved by technology

This kind of line-level correlation cannot be fully considered, and the correlation between trajectory points or lane line points cannot further improve the accuracy of trajectory prediction.

Images