Intelligent Analysis of the Results of Barrier Fence Monitoring

Abstract

The article analyzes the results of monitoring the barrier fence installed on the road, obtained using the complex multi-criteria assessment of the operational condition of the street and road network "ADS-MADI". The capabilities of data mining, which allows making informed decisions on the repair or replacement of blocks and assessing the dynamics of changes in their condition in the future, due to the identification of hidden patterns, were evaluated. The authors classified the detected defects into four categories. Using a multiple linear regression model, an analysis of the reasons for assigning defects to a particular group was performed. The results of factor analysis, which allows evaluating the relationship between the number of defects of each type, are presented. It is concluded that it is possible to selectively diagnose some sections and calculate the remaining sections of the barrier fence by multiple regression. The analysis results of the defects distribution of each category along the road are presented and their interpretation is given.