Abstract
Securing IEEE 802.15.4 wireless networks is one of the key challenges in the development of the Internet of Things (IoT). Given the limited computational resources of IoT devices, traditional attack detection methods based on cryptographic mechanisms and deterministic thresholds do not always provide a sufficient level of protection. In this paper, we propose a novel method for adaptive network traffic monitoring that combines a modified Z-score with sample size consideration and an adaptive Bayesian classifier with dynamic attack probability adjustment. Experimental testing on data generated in an NS-3 environment shows that the proposed method outperforms traditional approaches in terms of attack detection accuracy, reducing the false positive rate from 10.9% to 3.8%. The hybrid model provides 94.7% classification accuracy and 91.8% attack detection completeness, which is 6.3% higher than the standard Bayesian classifier. The obtained results demonstrate the promising use of the proposed method in real-time systems for monitoring the security of IoT networks. The developed approach allows adapting to the changing network environment, reducing the influence of random fluctuations, which makes it an effective solution for protecting low-power networks.