Optimized DDoS Detection in Smart Homes Using EPSO and Recurrent Transformer Networks

The risk of Distributed Denial-of-Service (DDoS) attacks on smart home systems is increasing due to the advanced nature of network complexity and the rise in the application of encrypted information. To improve DDoS detection, this research suggests using a new framework that combines an Enhanced Particle Swarm Optimization (EPSO) with a Recurrent Transformer Network (RTN). The collection and examination of network data is done at the packet level as well as the flow level. Through analyzing this information, the EPSO algorithm can detect important patterns even when they are in low volume, while the RTN can track temporal patterns even in encrypted communication. Experimental evaluations demonstrate the framework’s superiority over traditional methods, achieving higher accuracy, reduced false alarms, and improved smart home security against DDoS threats. Where we achieved an accuracy of 98%, recall of 99%, precision of 96%, F1-score of 97%, and an AUC of 99%.

DDoS Attacks, IoT Networks, Attack Detection, Encrypted Data, Smart Home Networks, Traffic Analysis

10.52113/2/12.01.2025/117-131