Artificial Intelligence Techniques for Hybrid Transformer based Gated Graph Attention Capsule Network Design for Preventing Attack in Radar Target Detection: Trends and Challenges

Abstract

Radar target detection plays a vital role in surveillance, defence, and autonomous sensing systems but is often affected by noise, clutter, electronic jamming, and adversarial attacks that reduce detection accuracy. Artificial intelligence, particularly deep learning, has emerged as an effective solution for improving radar target detection under challenging environments. This review examines recent advances in hybrid transformer-based gated graph attention capsule networks for robust radar target detection. Transformer models capture global contextual dependencies through self-attention, while graph attention networks model relationships among radar nodes and multi-sensor systems. Capsule networks preserve hierarchical spatial features, enhancing robustness against adversarial attacks and signal distortions. Their integration enables superior feature extraction, multi-dimensional feature fusion, and improved detection performance, especially in low signal-to-noise ratio conditions. The review also discusses current trends, including attention-based learning, graph neural networks, and hybrid deep learning architectures for intelligent radar systems. Despite significant progress, challenges such as computational complexity, model scalability, and real-time implementation remain. Future research should focus on lightweight, explainable, and energy-efficient hybrid AI models capable of providing secure, adaptive, and highly reliable radar target detection for next-generation surveillance and defence applications.