Method, device and computer-readable storage medium for speed prediction

Abstract

The embodiment of the present invention discloses a speed prediction method, device, and computer-readable storage medium. According to the trajectory data within a preset time period, the initial speed vector of the target road network is calculated; The convolutional neural network performs convolution processing on the initial velocity vector and the adjacent road section matrix to obtain the corresponding feature matrix; wherein, the feature matrix includes the corresponding feature vectors of each road section; each feature vector is converted into a time series, And the pre-trained long-short-term memory network is used to process the time series to obtain the target speed matrix corresponding to the target road network. Since the eigenvector is a vector obtained by considering the topological structure of the road network, it can effectively avoid the influence of the wrong space dynamic evolution feature on its accuracy. The target velocity vector predicted according to the feature vector is made more accurate, and the accuracy of velocity prediction is effectively improved.

Description

technical field [0001] The invention relates to the technical field of urban traffic, in particular to a speed prediction method, device and computer-readable storage medium. Background technique [0002] With the development of social economy, the number of private cars is increasing year by year, and the speed of road reconstruction and expansion in some big cities can no longer meet the requirements of the growth of motor vehicles. The resulting urban traffic congestion, traffic accidents and other problems have become a problem. Important issues of urban development and resident mobility. For this reason, Intelligent Transport Systems (Intelligent Transport Systems, ITS) has become a research hotspot as an effective solution. The traffic system is a time-varying, huge and complex system with human participation. How to accurately predict the ever-changing urban road traffic speed is one of the core issues of ITS. [0003] Various subsystems of ITS, such as timely adjus...

AI Technical Summary

Problems solved by technology

The traditional CNN can only learn the adjacent units in the matrix, but the city-level space-time velocity matrix cannot guarantee that the adjacent road sections are in adjacent rows. The above-mentioned method of using CNN to predict the traffic speed of the city does not take this into account, so It will learn some wrong spatial dynamic evolution features and affect its prediction accuracy

Images