Research
Themes
Intrusion detection. Models that detect attacks on databases, IoT networks, and IoMT devices. Methods include density-based clustering with ensemble learning, parallel metric learning, balanced ensembles for imbalanced traffic, and quantization-aware knowledge distillation for lightweight detection in SDN-IoT settings.
Fraud detection. Detection of fraud in mobile telecommunication networks, automobile insurance claims, credit card transactions, and online auctions. Methods include fuzzy clustering, support vector machines, and deep learning. A recurring topic is the effect of class imbalance on detection accuracy.
Distributed and secure learning. Federated learning, blockchain-based anomaly detection, and multi-agent coordination, mostly with R. C. Shit. Applications include urban traffic optimization, autonomous vehicle networks, and UAV collision avoidance. Preprints are listed on the Publications page.
Digital watermarking. Blind and perceptually guided watermarking of images using transform-domain embedding and metaheuristic optimization, with S. N. Bal.
Funded projects
- HealthWatch: Intrusion Detection in IoMT based Medical Devices. Principal Investigator. Mukhyamantri Research Innovation for Extramural Research Funding (MRIP-2024), Government of Odisha.
- Designing of Advance Intelligent Data driven framework for Mining Socioeconomic Factors Influencing Tribal Girl's Education in Rural Mayurbhanj District of Northern Odisha. Co-Principal Investigator. University Intramural Research Funded Project (UFIRP-2024), Maharaja Sriram Chandra Bhanja Deo University.
Patents
Six patents and patent applications are listed on the Publications page.