MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641096079 A) filed by Dhanavelu Anandan; Mr. A. Bharathidhasan; Dr. M. Rajendiran; and Mr. R. Palraj on August 08, 2026, for A Hybrid Cnn-Lstm Based Explainable Framework For Dos And Ddos Attack Detection.
Inventors include Dhanavelu Anandan; Mr. A. Bharathidhasan; Dr. M. Rajendiran; and Mr. R. Palraj.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: Abstract— Denial-of-Service (DoS) and Distributed Denial-of-Service (DDoS) attacks continue to pose significant challenges due to the rapid expansion of internet- based and cloud infrastructures. These attacks typically disrupt network availability by overwhelming systems with excessive traffic, leading to service degradation or complete failure. As network architectures grow more complex and dynamic, detecting such attacks in a timely and accurate manner becomes increasingly difficult. Traditional Intrusion Detection Systems (IDS) largely rely on predefined rules and signatures, making them effective against known threats but limited in identifying novel or evolving attack patterns. To overcome these limitations, recent research has shifted toward machine learning and deep learning techniques, which enable automated detection by analysing network traffic behaviour. Among these approaches, Convolutional Neural Networks(CNN)are effective in capturing spatial relationships within traffic data, while Long Short-Term Memory(LSTM)networks are well-suited for modelling temporal dependencies. Combiningthesemodelsallowsformorecomprehensive detection of complex attack patterns in dynamic environments. Current studies increasingly emphasize hybrid CNN-LSTM frameworks, along with efforts to improve model interpretability through explainable AI. This work therefore proposes a hybrid deep learning approach for enhanced DoS/DDoS detection.
Disclaimer: Curated by HT Syndication.
