MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202641109254 A) filed by Dr. Kavitha Nallamothu on September 11, 2026, for Hybrid Machine Learning Model For Efficient Botnet Attack Detection In Iot Environment.
Inventors include Dr. Kavitha Nallamothu; Sivaram Prasad Nalluri; Palem Naresh; Malla Sowmya; Kamadi Venkata Satya Ram Prasad Varma; Dr. Y. V. R. Naga Pawan; Dr. U. Mohan Srinivas; and M. Bhargavi.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: The proliferation of Internet of Things (IoT) devices has created a vast and vulnerable attack surface. These devices are highly susceptible to botnet attacks due to their inherent constraints in limited memory and processing power, which renders traditional antivirus solutions ineffective. This vulnerability poses a significant security risk, as compromised devices can be marshaled for large-scale Distributed Denial-of-Service (DDoS) attacks. This project proposes a novel Hybrid Machine Learning Model designed to substantially improve attack detection accuracy within resource-constrained IoT environments. The proposed solution is a stacked "ACLR" architecture that combines an ANN, CNN, LSTM, and RNN. This hybrid approach leverages the respective strengths of each model to effectively capture both spatial and complex temporal patterns inherent in network traffic data. Trained and validated on the benchmark UNSW15 dateset, the proposed stacked model achieved an impressive 99.86% accuracy, significantly outperforming the individual models. To further refine its precision, an enhanced version integrates an Attention Mechanism. This layer enables the model to dynamically weigh and focus on the most relevant features of the data, boosting the final accuracy to 99.91%. A comprehensive evaluation confirmed the model's robustness using standard performance metrics, including accuracy, precision, recall, F1-score, and ROC analysis. For practical application, the system is implemented using Jupyter Notebook for training and a Flask web framework for deployment. This provides a user-friendly web interface, allowing operators to upload test data and receive real-time botnet detection results. This approach ensures a lightweight, efficient, and highly accurate solution for securing modern IoT networks against sophisticated botnet threats.
Disclaimer: Curated by HT Syndication.
