Pushing the Boundaries of Urban Wireless Networks: A Semantic-Aware Intelligent Network System (SA-INS) for Data-Centric Adaptive Spectrum Management

Abstract

Urban wireless networks are increasingly challenged by the surge in connected devices and data-intensive applications, which exacerbate spectrum scarcity and resource allocation complexities. This paper introduces the semantic-aware intelligent network system (SA-INS), a data-centric framework that redefines resource management through a semantic-aware prioritization approach. By evaluating the contextual content of data, SA-INS ensures optimal information is allocated, surpassing conventional prioritization methods. Additionally, it employs a Q-learning-based algorithm tailored to the unique requirements of semantic-aware communication systems. These enhancements address scalability, multi-objective optimization, and efficient resource allocation. To improve spectrum efficiency in densely connected environments, advanced techniques from 5G and 6G networks are integrated, including in-band full-duplex (IBFD), non-orthogonal multiple access (NOMA), and a water-filling power allocation algorithm. To further refine resource distribution, network slicing is implemented to create virtual networks tailored for specific data types, ensuring efficient quality of service (QoS) provisioning. A multi-objective optimization framework, using the non-dominated sorting genetic algorithm II (NSGA-II), is also incorporated to achieve balanced optimization of conflicting objectives. A comprehensive mathematical framework is developed to rigorously model and analyze the performance of SA-INS, providing deeper insights into system dynamics and trade-offs. Finally, simulation results validate the effectiveness of the proposed system, demonstrating significant improvements in spectrum efficiency, latency, and packet loss rates (PLR) compared to benchmarks.