A method and system for predicting traffic flow at expressway toll stations

Abstract

The invention belongs to the field of traffic data processing and discloses a method and system for predicting traffic flow at a highway toll gate. The method comprises steps of: performing clusteringby using k points in the space as a center by using a clustering method, classifying the closest objects, gradually updating the values of respective clustering centers by using an iterative method until an optimal clustering result is obtained. The method for predicting the traffic flow at a highway toll gate also includes processing the data by using association rules and calling an arules packet. The method solves a problem that most modeling processes are similar to a traditional time sequence model, cannot dynamically adjust model parameters, cause reduced sequence stability due to overlarge sample capacity, and are complicated in modeling process.

Description

technical field [0001] The invention belongs to the field of traffic data processing, and in particular relates to a method and system for predicting traffic flow at expressway toll stations. Background technique [0002] Traffic flow forecasting can effectively alleviate traffic congestion, reduce the incidence of accidents, and provide travelers with a comfortable and safe traffic environment. Road traffic changes are a real-time, non-linear, non-stationary random process. The shorter the statistical period, the stronger the randomness and uncertainty of short-term traffic flow changes. Since the intersection is an important part of urban traffic, the traffic situation is complex and there are many interference factors, the problem of insufficient traffic capacity has become the bottleneck of the traffic network. Therefore, research on short-term traffic flow forecasting at intersections is a key issue in realizing urban traffic intelligence. [0003] Real-time and accur...

AI Technical Summary

Problems solved by technology

[0005] Most of the modeling process is similar to the traditional time series model, the time interval is large, the forecast data is not accurate, there are model parameters that cannot be adjusted dynamically, the sample size is too large to weaken the stationarity of the sequence, and the modeling process is complicated. When there are multiple periodic components at the same time, it is difficult for the autoregressive moving average model with a lower order to reflect multiple periods at the same time, and the autoregressive moving average model loses sample points, which will cause severe traffic flow time series. The problem

[0006] The reason why technical problems cannot be effectively solved is that there are many factors affecting traffic flow in different regions, and it is not accurate to build a model from the traffic data itself to make predictions. However, the actual traffic conditions still have certain contingencies, and technical problems can only be solved as much as possible. Reduce the error from the actual

[0007] The difficulty of solving lies in: the uncertainty and specific degree of influence on factors affecting traffic flow data, involving a large number of data types and different data formats, as well as subjective factors and travel behavior characteristics of specific regional populations

At the same time, traffic flow has complex nonlinear characteristics, and the traffic flow time series that needs to be analyzed are mostly nonlinear and non-stationary data. Various classical time series analysis methods will have accuracy when encountering nonlinear and non-stationary data. Insufficient defects, or the obtained results do not have a clear meaning of the actual traffic situation. Therefore, it is necessary to use a new method to analyze and study the nonlinear and non-stationary data in the traffic flow fluctuation, and consider the special worth influencing factors.

Images