Data Control in Distributed Self-organizing Sensor Network Under Speciffic Deployment Condition

Abstract

A wireless sensor mesh network can be deployed in special conditions where stable GSM, Wi-Fi, and other coverage are absent. At the same time, there is also a risk of malicious interference with the transmitted information. A popular cognitive radio (CR) communication device with spread spectrum technology is also susceptible to radio jamming. Recent practical knowledge regarding radio jamming, radio reconnaissance, and electronic warfare indicates that radio jamming is effective but has certain limitations in terms of working distance and energy consumption. A foundational practical test under conditions approximating special deployment circumstances revealed that typical radio jamming schemes operate at distances up to 1–1.3 km when applied to cognitive radio (CR) devices. These data, overall, are corroborated by the practical application of networks of this kind in specialized deployment conditions. Additionally, Wireless Sensor Networks (WSMN) are vulnerable to man-in-the-middle attacks, which are challenging to identify. The utilization of machine learning methods and the XGBoost algorithm for analyzing the content of sensor network data frames provide close to 100% probability of detecting data substitution within a frame. The use of this mentioned method facilitates rapid training, both based on synthetic data and real-world data within the system and does not require significant computational resources.