A Hybrid AI-Driven Cybersecurity Framework for Secure Communication Systems in ECE Networks

Abstract

The rapid growth of interconnected communication systems in Electronic and Communication Engineering (ECE) networks has increased the risk of cyber threats, making secure data transmission a critical challenge. Traditional intrusion detection approaches often struggle to handle dynamic attack patterns and complex network behaviour, resulting in limited accuracy and high false alarm rates. The objective of this study is to develop an efficient and adaptive cybersecurity framework capable of improving intrusion detection performance in modern communication environments.The proposed work introduces a hybrid AI-driven cybersecurity framework that integrates multiple learning techniques to analyse network traffic effectively. The system utilizes a structured pipeline involving data preprocessing, feature engineering, and hybrid model design combining spatial and temporal learning mechanisms. The KDD Cup 1999 dataset is used for training and evaluation, with features categorized into basic, content-based, time-based, and host-based attributes. The model is trained using optimized parameters and evaluated through standard performance metrics to ensure reliability and reproducibility.Experimental results demonstrate that the proposed framework achieves an accuracy of 95.2%, precision of 94.1%, recall of 93.6%, and an AUC score of 0.95, outperforming traditional machine learning and standalone deep learning models. The results also indicate stable convergence, reduced loss, and improved detection capability across multiple attack categories.The study concludes that the hybrid approach significantly enhances intrusion detection performance while supporting real-time adaptability and scalability. The framework provides a practical solution for securing ECE  communication systems and has strong potential for deployment in real-world applications such as IoT, smart grids, and industrial networks