Models and Methods of Intrusion Detection in Wireless Sensor Networks of the Tactical Command and Control of Troops

Wireless sensor networks are an important element of modern military operations, providing real-time monitoring and data transmission. However, these networks are vulnerable to both physical and cyber attacks due to limited resources, lack of physical control over the sensors, and challenges associated with using wireless communication channels. The aim of the article is to conduct a comparative analysis of models and methods for intrusion detection in tactical command-level wireless sensor networks. The analysis covers centralized and decentralized security management approaches with a focus on detection models based on signatures, anomalies, and specifications. The article also explores the potential of using hybrid methods that combine the advantages of the aforementioned approaches. Publicly available datasets (KDD, NSL-KDD, WSN-DS) and synthetic datasets generated using network simulators were used to compare the effectiveness of the models. The results show that centralized models are more effective for small networks but create a load on the base station, which can cause delays in attack detection. Decentralized models reduce the load and improve the speed of response to attacks, but they also have their drawbacks. The article notes that none of the existing methods provide complete protection, so a combination of approaches is the most effective solution. Anomaly-based intrusion detection models and methods are classified according to their functional capabilities: statistics-based, data mining-based, machine learning-based, and artificial intelligence-based. The use of artificial neural networks and machine learning significantly improves the accuracy of anomaly detection, but such systems require large computational resources and are complex to configure. The main analytical conclusion of the article is the need to create a hybrid intrusion detection system using artificial neural networks and machine learning, which combines centralized and decentralized methods while considering specific threats to tactical command-level wireless sensor networks. Future research should focus on developing a functional model of an intrusion detection system for the security subsystem in tactical command-level wireless sensor networks.