Introducing Innovations to Improve the Automatic Rolling Stock Identification System

The article is devoted to the analysis and development of automatic rolling stock identification systems using modern technologies. The paper analyzes the existing identification methods, considers the types of automated systems and their main tasks. Particular attention is paid to video surveillance and radio frequency identification (RFID) technologies, as well as the specifics of using Amtech, Dynicom, and the Palm system. The design of a microprocessor-based system for automatic identification of rolling stock is also discussed in detail. The principles of operation, system organization, logic, and stages of the identification algorithm are considered. An algorithm for counting the number of cars and analyzing the functioning of the DPD-01, wheel set passage sensors is proposed. The development of an optical identification subsystem based on neural networks is also considered. The structure and mechanism of functioning, algorithms for recognizing the numbers of rolling stock, as well as the use of artificial intelligence are investigated. The features of neural network training, their topology and optimization are studied. Technical modeling of the optical identification system was performed. The research emphasizes the importance of integrating advanced technologies to enhance the reliability and accuracy of identification systems. It highlights the challenges faced in real-time identification and suggests solutions based on sensor fusion and AIdriven decision-making. Practical implementation issues and potential improvements in data processing speed and recognition accuracy are discussed. The study concludes that the proposed system can significantly improve operational safety and logistical efficiency in railway transport. The obtained results can be used to improve the efficiency of automation of rolling stock identification in railway transport.