Traffic control subregion clustering and dividing method based on multi-source data fusion and SNMF

Abstract

The invention discloses a traffic control subregion clustering and dividing method based on multi-source data fusion and SNMF (Symmetric Non-negative Matrix Factorization). The method comprises the steps of: firstly, selecting two characteristic parameter data for correlation analysis according to an actual traffic environment, and performing linear weighted data fusion on the two characteristic parameter data on the basis of data normalization processing to obtain a new combined characteristic parameter; then, according to the combined characteristic parameter, adopting a hierarchical clustering algorithm to generate a 'snake' array capable of representing the surrounding traffic flow situation for all road sections; and finally, calculating the similarity matrix of all road sections, and performing SNMF optimization solution to obtain a final traffic control subregion clustering and dividing result. The method is suitable for medium and small-sized urban traffic networks, the traffic control subregion division result which is more in line with the actual traffic condition can be obtained based on the combination characteristic parameters of multi-source data fusion, and meanwhile, the SNMF has good decomposition characteristics, so that the defects of the existing clustering division method are avoided.

Description

technical field [0001] The invention relates to a traffic control sub-area division method for intelligent traffic signal control. The traffic control sub-area is used for arterial coordinated control and regional coordinated control of urban traffic signals. Background technique [0002] In urban traffic signal control, coordinated control can effectively improve the traffic efficiency of the entire system, reduce parking delays and travel time per vehicle. However, in the urban traffic road network, there are different degrees of differences in the dynamic traffic flow characteristics of each intersection and road section. If it is used as the same area to implement a unified control strategy, it will not achieve a good control effect, or even Aggravate congestion or cause traffic accidents. The division of traffic control sub-areas is mainly to divide adjacent intersections or road sections into several traffic control sub-areas for coordinated control. It is the coordin...

AI Technical Summary

Problems solved by technology

[0005] At present, in practical applications, the existing traffic control sub-area clustering division methods have the following main problems: 1) Most methods use characteristic parameters from a single source, which cannot accurately and comprehensively represent the traffic flow situation, resulting in inaccurate division results. It meets the actual needs; 2) Although a few methods use multi-source characteristic parameters, there are problems such as data collection difficulties or large collection errors; 3) Traditional clustering methods (that is, the first type of methods) have many deficiencies

K-means clustering has problems such as difficulty in selecting the initial cluster center and inaccurate division, while the effect of spectral clustering is overly dependent on the eigenvalues ​​of the Laplacian matrix, and non-negative matrix decomposition requires the data to have a good linear structure;4 ) Other clustering methods (that is, the second method) also have certain defects

Heuristic algorithms often can only obtain locally optimal subregion division results. Although modeling optimization methods can obtain optimal results, the calculation is very time-consuming

Images