AI-Driven Signal Processing and Network Management for Next-Generation Communications

Abstract

The acceleration of wireless technologies from 5G toward 6G has intensified demands for intelligent, adaptive, and secure communication infrastructures. This paper provides a comprehensive synthesis of artificial intelligence (AI) methodologies that advance signal processing and network management across next-generation communication systems. First, we examine AI-driven intelligent signal processing, including deep learning techniques for modulation recognition, waveform generation, channel estimation, and interference mitigation, emphasizing their applicability under heterogeneous, high-dimensional, and noisy environments. Cognitive radio and dynamic spectrum access mechanisms are analysed with a focus on learning-based spectrum sensing and adaptive policy selection, enabling efficient exploitation of scarce spectral resources. In parallel, we investigate IoT communication constraints—energy, latency, jamming resilience—and how AI optimizes scheduling, traffic adaptation, and resource usage within ultra-low-power environments.

At the network level, we discuss the integration of AI with Open RAN architectures, virtualization frameworks (NFV/SDN), network slicing, and distributed orchestration for 5G/6G. Special emphasis is placed on AI-supported resource allocation, multi-agent reinforcement learning, edge intelligence, and federated learning for collaborative, privacy-preserving management across large-scale, heterogeneous networks. Security challenges—including anomaly detection, adversarial robustness, and physical-layer protection—are evaluated within the broader context of AI-enabled network defence. Finally, we highlight benchmarking limitations and the need for standardized datasets, reproducibility protocols, and deployment-ready evaluation criteria. Collectively, the findings underscore AI as a foundational enabler of resilient, efficient, and scalable communication systems for emerging 6G and beyond.