Secure 6G Mobile Devices Using Deep Kronecker Neural Networks and Hybrid Optimization: A Review

Abstract

The emergence of sixth-generation (6G) communication networks introduces unprecedented security challenges, particularly for mobile devices operating in distributed and resource-constrained environments. Among these threats, side-channel attacks (SCAs) pose a significant risk by exploiting indirect physical leakages such as power consumption, electromagnetic emissions, and timing information to compromise cryptographic systems. Traditional security mechanisms are insufficient to counter these advanced attacks, necessitating the adoption of artificial intelligence (AI)-based solutions. This paper presents a comprehensive review of AI techniques for secure 6G mobile devices, focusing on Deep Kronecker Neural Networks (DKNN) optimized with Hybrid Cat Hunting Optimization (HCHO). The study analyzes recent developments in deep learning, reinforcement learning, federated learning, and hybrid optimization techniques for detecting and mitigating SCAs. The findings indicate that hybrid AI models significantly improve detection accuracy, computational efficiency, and adaptability. DKNN reduces model complexity through Kronecker factorization, while HCHO enhances convergence and parameter optimization. The paper further explores emerging trends such as AI-native security architectures, edge intelligence, and quantum-enhanced security. Finally, key challenges including computational overhead, energy constraints, adversarial vulnerabilities, and lack of standardization are discussed, providing directions for future research.