Prediction method of urban expressway travel time based on spatio-temporal grid data of floating cars

Abstract

The invention relates to a city expressway travel time prediction method based on floating car space-time grid data, and belongs to the field of intelligent traffic. According to the invention, floating car data is firstly processed to obtain a time-space velocity matrix; instant travel time and real travel time of passing through a whole target route for each given departure time are calculated; secondly, a historical database is established and two kinds of historical databases of working days and non-working days are established; a prediction model is established, and the time-varying characteristics of traffic are extracted from the space-time velocity matrix by using a gray level co-occurrence matrix in establishment; and similar historical traffic states are selected from the historical database, and then secondary matching between the instant travel time of to-be predicted traffic states and historical traffic states are performed to conduct weight allocation to obtain final predicted travel time. The city expressway travel time prediction method of the invention fully exploits the characteristics of the historical data, and the model is simple and efficient; no long-term training process, prior knowledge or parameter identification is needed; and the method has high fault-tolerance and robustness and high stability.

Description

technical field [0001] The invention belongs to the field of intelligent transportation, and accurately grasps the time-space characteristics of urban expressway traffic flow based on the space-time grid data of floating vehicles, and accurately predicts the travel time of the urban expressway. Background technique [0002] As an important indicator for evaluating traffic conditions, travel time information has become a key component of advanced traveler information systems and advanced road traffic management systems. Real-time and accurate travel time information release is essential for refined traffic management and improved travel services. It has important theoretical research value and practical significance. [0003] Travel time prediction is a long-term research hotspot in the field of ITS (Intelligent Transport System, Intelligent Transportation System), and various prediction models have emerged in the past few years. Most of the previous prediction models were b...

AI Technical Summary

Problems solved by technology

[0004] Generally speaking, travel time prediction models mainly include model-driven methods and data-driven methods. Model-driven methods include macro-traffic flow models, time series methods, and Kalman filter methods. The influence of interference factors is large and the uncertainty is strong; data-driven methods include neural network methods, support vector machines, K-nearest neighbor methods, etc. This type of method does not require prior knowledge and parameter identification, and has strong fault tolerance and robustness characteristics, but there are certain requirements for the amount of historical data

Considering the highly time-varying characteristics and non-linear characteristics of urban expressway traffic conditions, it poses a great challenge to the accurate prediction of travel time

Images