Multimodal intelligent decision support system for improving industrial safety at oil and gas facilities, Мультимодальная интеллектуальная система поддержки принятия решений для повышения промышленной безопасности объектов нефтегазовой отрасли

<jats:p>Relevance. Ensuring industrial safety at oil and gas facilities requires simultaneous monitoring of process parameters, equipment status, personnel actions, and early warning signs of emergency situations. In practice, this data often remains dispersed among independent subsystems: video surveillance, fire alarms, gas analysis, access control systems, and industrial automation. As a result, the operator receives not a unified risk picture, but a set of disparate signals that must be quickly correlated by time, location, and meaning. Aim. To substantiate an approach to creating a local multimodal intelligent decision support system that combines video, audio and sensory data. Methods. The methodological basis of the work is system analysis, conceptual modeling, multimodal data fusion, and microservice architecture principles. Results. The authors have proposed a system model, including a data collection layer, preprocessing, computer vision, audio analytics, and sensor diagnostics modules, a RabbitMQ message bus, a correlation aggregator, a decision support module, and an operator interface. It is demonstrated that cross-verification of events across multiple independent sources can improve the reliability of alarm messages and reduce operator workload. The scientific significance of this work lies in the formalization of an approach to the joint interpretation of heterogeneous data, while its practical significance lies in the feasibility of applying a local architecture to critical information infrastructure facilities. For citation: Chumakov K.V., Zhdanova A.O. Multimodal intelligent decision support system for improving industrial safety at oil and gas facilities. Bulletin of the Tomsk Polytechnic University. Geo Assets Engineering, 2026, vol. 337, no. 8, pp. 184–199. http://doi.org/10.18799/24131830/2026/8/5806</jats:p>