Vehicle trajectory feature extraction method based on directed graph structure and LSTM

Abstract

The invention discloses a vehicle trajectory feature extraction method based on a directed graph structure and an LSTM, and relates to the technical field of urban traffic. According to the vehicle trajectory feature extraction method, vehicle trajectory point data and mature road data are effectively combined to realize standardized processing of the vehicle trajectory point data. A sequence formed by corresponding local trajectory features is selected and constructed from four aspects including time, space, motion and environment to describe the motion behavior of the trajectory, the features of multiple aspects of the trajectory are fully fused, meanwhile, the space-time dependency relationship of the trajectory and the quantitative features of the trajectory are comprehensively considered, the deep-level expression of the trajectory is learned by using a deep learning method, and finally, the trajectory features are described in a vector form. Standardized feature data are provided for vehicle trajectory classification or prediction, and the method has certain universality.

Description

technical field [0001] The invention relates to the technical field of urban transportation, in particular to a vehicle trajectory feature extraction method based on a directed graph structure and LSTM. Background technique [0002] As a kind of dynamic geographic data, trajectory modeling and prediction research has always been a hot topic in geography and GIS; trajectory modeling is also one of the core issues of spatio-temporal big data mining and modeling. It is of great significance to study the problems and explore unmanned driving technology. [0003] Most of the existing trajectory feature representations are artificial features, and the consideration of the diversity of the characteristics of the trajectory itself is relatively limited in the acquisition, and the consideration of the trajectory space sequence features mainly depends on the sequence of trajectory coordinate points. Contents of the invention [0004] Aiming at the deficiencies in the prior art, the...

AI Technical Summary

Problems solved by technology

[0003] Most of the existing trajectory feature representations are artificial features, and the consideration of the diversity of the characteristics of the trajectory itself is relatively limited in the acquisition, and the consideration of the trajectory space sequence features mainly depends on the sequence of trajectory coordinate points

Images