Multi-Agent Architecture for Intelligent Rolling Stock Diagnostics: Concept and Prototyping
Year: 2025
Type: article
Abstract: The paper considers the development of an intelligent diagnostics and recovery system for rolling stock information management system operability, based on a multi-agent architecture. Introduction: the growing complexity of rolling stock information management systems (IMS) demands for greater reliability and more advanced intelligent diagnostic and recovery systems. Traditional centralized methods can become inefficient under certain conditions. Purpose: to develop the concept of an intelligent system for diagnosing and restoring the operability of rolling stock IMS. This system is based on a multi-agent architecture that can be implemented in existing transport infrastructure. Methods: the proposed approach involves the implementation of a specialized task allocation system, wherein each agent within the system is assigned a specific task. These tasks encompass the collection of telemetric information, the analysis and filtration of data, the formation of diagnostic indicators, and the undertaking of decisions concerning the technical condition of the system. The system’s architecture facilitates multi-level organization and interaction protocols between the agents. Results: the development of an architectural model and a prototype concept has been undertaken for the purpose of demonstrating the functionality of a distributed intelligent system. The presentation has focused on the key components, interaction interface, and hardware and software implementations. Practical significance: the developed approach has the potential to enhance the stability and autonomy of transport systems in real operating conditions. Discussion: the potential for scaling the prototype and integrating it into existing IMS is a subject of discussion.
Authors: Anatoly Baranovsky, Daniil Bobrikov
Institution: Petersburg State Transport University
Language: English
Field: Computer Science
Domain: Physical Sciences
Open Access status: bronze
OpenAlex ID: W4414579907