APPLICATION OF CLUSTER ANALYSIS OF INTRSECTIONS FOR ENHANCEMENT OF THE EFFICIENCY OF THE WORK OF TRAFFIC LIGHT OBJECTS

  • Я. В. Баштинська O. M. Beketov National University of Urban Economy in Kharkiv
  • В. С. Величко O. M. Beketov National University of Urban Economy in Kharkiv
  • Н. І. Кульбашна O. M. Beketov National University of Urban Economy in Kharkiv http://orcid.org/0000-0003-1551-1500

Abstract

In the presented article a cluster analysis of intersections that takes into account the conditions of motion and its environment on the choice of systems of traffic-light control, modes and algorithms of work is proposed for the purpose of improving the efficiency of the performance of traffic lights objects. For the classification of regulated intersections, a set of classification characteristics is used: the number of lanes, the gravity potential of the intersection, the presence of stops and public transport lines, the intensity of pedestrian traffic, the number of accidents, the complexity of the regulated intersection.

The results of the classification of regulated crossroads on the selected object are presented – these are regulated crossroads of Gagarin avenu of Kharkiv. A statistical cluster analysis is used to classify regulated crossroads. Based on the methodology, methods of analysis and classification, all regulated crossroads were divided into four classes. Classification functions for a defined class, the analysis of which showed that the weighting classification features are the potential of gravity of the crossroads (the location of infrastructure around), the number of traffic accidents and the complexity of the regulated crossroads, which is determined by the scheme of vehicle disperse, are obtained. Up to the first class crossroads belong less complex ones, to the second one – very complex. It is proposed to take into account such classification when choosing means and algorithms for traffic control.

In further research, variants of neural network training for neurocontrollers can be offered, which will be based not only on one input parameter – the intensity of traffic flows, but also take into account others ones, which are used in cluster analysis. It is envisaged,  that the neural network training to control the traffic light object of a simple and very complex crossroads should differ due to the action of different from the point of view of safety factors.

Keywords: intersection, traffic light control, traffic conditions, cluster analysis, control means

Author Biographies

Я. В. Баштинська, O. M. Beketov National University of Urban Economy in Kharkiv

магістр

В. С. Величко, O. M. Beketov National University of Urban Economy in Kharkiv

студент

Н. І. Кульбашна, O. M. Beketov National University of Urban Economy in Kharkiv

кандидат технічних наук., старший викладач

Published
2018-12-19
How to Cite
БаштинськаЯ. В., ВеличкоВ. С., & КульбашнаН. І. (2018). APPLICATION OF CLUSTER ANALYSIS OF INTRSECTIONS FOR ENHANCEMENT OF THE EFFICIENCY OF THE WORK OF TRAFFIC LIGHT OBJECTS. Municipal Economy of Cities, 7(146), 34-39. Retrieved from https://khg.kname.edu.ua/index.php/khg/article/view/5294