Scientific journal
Scientific Review. Technical science
ISSN 2500-0799
ПИ №ФС77-57440

Personal portfolio

DIAGNOSIS OF THE TRANSPORT INFRASTRUCTURE USING NEURAL NETWORKS

Sidorov A.V. 1 Mikheev S.V. 1 Osmushin A.A. 1
1 Samara State Aerospace University
Consider of neural networks for solving the tasks of monitoring, control and diagnostics of transport infrastructure. The main problems of studying the state of transport infrastructure and the necessity of intellectualization of information processing involving data mining techniques. To investigate the transport infrastructure presents a method of Fault Detection and Identification (FDI), and shows how its application regarding the problem. A comparative analysis of FDI method with the least squares method. For the implementation of algorithms for diagnosing the state of the transport infrastructure of selected neural network Kohonen self-organizing . The efficacy of the use of Kohonen neural network in comparison with the least squares method by automatically partitioning the diagnostic space features state of transport infrastructure to the class. Proved the possibility of solving the problem of different architectures of neural networks.