INTELLIGENT METHODS FOR DETECTION OF PERFORMANCE DISORDERS IN WIRELESS SENSOR NETWORKS

Abstract

This article explores advanced intelligent methods for detecting malfunctions in wireless sensor networks (WSNs), with a particular emphasis on enhancing operational reliability under conditions of limited resources. The study substantiates the viability of employing hybrid machine learning models that integrate traffic characteristics, topological information, and energy parameters to comprehensively assess the state of network nodes. A novel multi-level diagnostic system architecture is proposed, consisting of sequential stages including real-time monitoring, intelligent classification, and precise localization of network faults. This layered approach ensures improved adaptability and responsiveness in dynamic WSN environments. Simulation experiments were conducted using the AnyLogic and OMNeT++ platforms to evaluate and compare the effectiveness of conventional diagnostic methods versus the proposed intelligent approaches. The results clearly demonstrate the superiority of the intelligent models in terms of diagnostic accuracy, reduced fault detection latency, and optimized energy consumption. The proposed methodology significantly enhances the robustness and efficiency of WSN operations, making it particularly suitable for application in distributed Internet of Things (IoT) infrastructures and other systems where computational resources are constrained. These findings underscore the potential of intelligent diagnostic frameworks to ensure high data availability, prolong the operational lifetime of sensor nodes, and minimize the risk of information loss in critical network deployments.