Lipchanska O. The methods of data processing and data transmission for the locomotive driver's information support subsystem

The thesis solves the problem of developing methods for processing and transmitting data for the locomotive driver's information support subsystem. Based on the study of modern methods and means of processing and transmitting data on railway transport, it is established that there is a contradiction between the increase in the speed and intensity of railway traffic, the increase in the number of emergency areas, the complexity of the railway transport management system, the increase in the requirements for railway transport safety and the applied mathematical apparatus and technologies of coding / decoding. The choice of technology of mathematical formalization based on queuing networks and neural networks is justified. A conceptual model of the wireless segment of the locomotive driver's information support subsystem has been developed, which is built using 4G technology and takes into account the QPSK modulation type and the interleaving procedure for frame transmission to reduce the probability of erroneous reception while ensuring a high data rate. A mathematical model of the process of transmitting video data over a physical distributed uplink channel that takes into account the peculiarities of video data transmission in accordance with the LTE standard under dynamic conditions of the railway main line is developed and allows to estimate a wide range of probability-temporal characteristics of the video data transmission process in the locomotive driver's information support subsystem. A mathematical model of access control to the segment of a locomotive driver's information support subsystem has been developed, which takes into account the peculiarities of video data transmission in accordance with the LTE standard in the dynamic conditions of the railway line. The procedure for noiseless decoding of streaming video in the wireless 4G network segment of the locomotive driver’s information support subsystem is improved. It allows to increase efficiency in data transfer, to reduce the total number of calculations during decoding and to release the computing resources of the computerized system. An intelligent system for monitoring the state of dangerous sections of the railway using an adapted convolution neural network has been developed, which makes it possible to increase the reliability of the recognition of dangerous situations at railway crossings and inform the driver in the event of a critical situation. Practical recommendations as for the developed methods application are substantiated.