Citation

Attaullah, Hafiz Muhammad and Ehsan, Muhammad and Khan, Inam Ullah and Alam, Muhammad Mansoor and Mohd Su'ud, Mazliham (2026) Performance Benchmarking of ROBOTa Optimized Machine and Deep Learning Models for UAV IDS. In: Proceedings of the 3rd International Conference on Emerging Trends and Innovation (3rd ICETI). Springer Nature, pp. 817-834. ISBN 978-3-032-22425-5, 978-3-032-22426-2

Text Performance Benchmarking of ROBOTa Optimized Machine and Deep Learning Models for UAV IDS _ Springer Nature Link.pdf - Published Version Restricted to Repository staff only Download (739kB)

Abstract

The rapid proliferation of unmanned aerial vehicles (UAVs) has intensified the need for robust intrusion detection systems (IDS) capable of safeguarding UAV communication networks from increasingly sophisticated cyber threats. Traditional IDS frameworks often struggle to maintain high accuracy and low latency when confronted with complex, high-dimensional UAV traffic data. To address these challenges, this study presents a comprehensive comparative analysis of machine learning (ML) and deep learning (DL) models enhanced through ROBOTa-based feature optimization. The proposed framework employs the ROBOTa algorithm to select and refine the most discriminative features, thereby improving model efficiency and generalization. A diverse set of ML algorithms (SVM, Random Forest, XGBoost, KNN, Logistic Regression) and DL architectures (CNN, LSTM, BiLSTM, GRU) were trained and evaluated on benchmark UAV network datasets. Performance metrics–including accuracy, precision, recall, F1-score, AUC, and computational cost–were used to assess each model’s capability. Experimental results demonstrate that ROBOTa-optimized features significantly enhance detection performance across all models, ROBOTa with LSTM model achieving superior results, attaining an accuracy of 98.6% and a markedly reduced false alarm rate. This comparative evaluation highlights the efficacy of optimization-driven feature engineering in strengthening UAV network security and establishes ROBOTa with LSTM as a promising candidate for next-generation intelligent IDS frameworks.

Item Type:	Book Section
Uncontrolled Keywords:	Cyber-physical security, Deep learning
Subjects:	Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Divisions:	Faculty of Computing and Informatics (FCI)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	05 Oct 2026 03:31
Last Modified:	05 Oct 2026 03:31
URII:	http://shdl.mmu.edu.my/id/eprint/16869

Downloads

Edit (login required)