Humphrey Rutagemwa

Effective spectrum management is essential with the growing demand for wireless services and the increasing complexity of spectrum usage scenarios. However, regulatory authorities often encounter limitations in resource availability, posing challenges to their ability to efficiently manage the finite radio frequency spectrum. In response to these challenges, data-driven methodologies and the judicious use of artificial intelligence (AI) have emerged as promising approaches to enhance spectrum management practices. In this article, it is reasoned that the adoption of AI and data-driven approaches in spectrum management not only mitigates resource constraints but also enables regulators to automate routine tasks, streamline processes, and allocate their limited resources more effectively, allowing personnel to focus on higher-value activities.

Spectrum management and regulation are becoming more complex due to rapid technological advancements, increasing demand for spectrum, and the presence of diverse stakeholders with conflicting interests. In response, governments worldwide are increasingly interested in leveraging advanced technologies, such as Artificial Intelligence (AI), to enhance efficiency and optimize policy outcomes. This paper explores the application of large language models (LLMs), a subset of generative AI, to streamline tasks and improve decision-making in spectrum management and regulation. It examines the various roles of LLMs in this field and addresses associated challenges. Through empirical case studies and experimental findings, the paper demonstrates how LLMs can profoundly transform spectrum management and regulation practices. The study also offers insights into effectively integrating AI into regulatory frameworks, providing practical lessons and best practices for governmental AI initiatives.

As digital economies continue to grow and the demand for wireless communication intensify, efficient and effective spectrum management becomes increasingly vital. In response, regulatory bodies worldwide are interest in leveraging advanced technologies to enhance and expedite their spectrum management efforts. This paper investigates the application of large language models (LLMs) for regulatory processes and their potential to improve operational efficiencies and decision-making. In addition, this paper demonstrates how LLMs enhance and accelerate policy and regulation processes for information extraction and question answering; focusing on spectrum management use cases. The results obtained indicate a robust document processing alternative, suggesting a transformative impact on spectrum regulatory practices.

With the advent of sixth generation (6G) Internet of Things (IoT), integrated terrestrial network (TN) and non-terrestrial network (NTN) will play a vital role in enabling new applications and services. However, realizing the potential of 6G integrated TN-NTN requires addressing key challenges like intelligent and optimized control mechanisms for resource management, interference cancellation, and handover management. This paper explores the potential of large language models (LLMs) in intelligent network control for 6G integrated TN-NTN. LLMs can learn complex relationships and patterns from large-scale data, and then be fine-tuned on small labeled datasets, significantly reducing training time and cost. This study examines the main obstacles in the integration of 6G IoT and TN-NTN systems, and further discusses how intelligent control may effectively address those issues. Our suggested approach utilizes LLMs to create efficient anaptive control algorithms that can effectively handle the diverse, ever-changing, and decentralized characteristics of 6G integrated TN-NTN.

A common technique for improving throughput and increasing spectrum efficiency in wireless networks is the use of header compression. Removing static, implied or redundant protocol fields by synchronizing state information between a sender and receiver reduces per-packet overhead for long lived communication flows. Previous header compression work has focused primarily on single-hop unicast traffic. In this paper we describe a multi-hop multicast header compression (MMHC) scheme applicable to mobile ad-hoc networks. MMHC provides an end-to-end solution where multicast forwarders and receivers are synchronized simultaneously and not hop-by-hop as in previous schemes. Simulation results show that the reductions in per-transmission overhead achieved provide significant gains in both throughput and latency, especially when the average payload size of packets aggregated into the communication flow is small.

A framework for frequency adaptation is proposed to enable networks that coexist in uncoordinated spectrum bands to gradually change their operating frequencies to attain interference-tolerant conditions. It is shown that incorporating channel conditions and cross-layer information in making frequency adaptation decisions significantly increases spectrum utilization efficiency and reduces access delay.

This work studies the networking scheme of NATO's next generation narrow band waveform (NBWF) radios. The key techniques that effectively enable the multihop capabilities for the NBWF radios are described first. Then a network emulator is developed to evaluate two critical algorithms, the link metrics mapping algorithm and the relay node selection algorithm, employing realistic network scenario parameters. Performance results are illustrated and the options for dynamic algorithm adaptation are elaborated.

In today's network-centric battlefield, the ad hoc style, self-organizing networks play a more and more important role in the operation of mobile forces that are deployed quickly to meet the tactical demands. However, the unique features of the tactical scenario also pose significant challenges for networking, which makes the existing research results for the general mobile ad hoc networks (MANETs) difficult to be reused directly. In this paper, we focus on the connectivity issue of the mobile tactical networks (MTNs). To better describe the special formation-oriented mobility pattern of MTNs, a more realistic node-following mobility model is used, based on which the neighbor connectivity is obtained from both the geometrical and physical communication point of view for not only 1-D but also 2-D network scenarios. After that, we further evolve our recently proposed decomposition and recursion method to derive the end-to-end connectivity for some formations commonly used in the MTNs. All our analytical results are verified by extensive simulations. We believe that this work is able to shed new lights on the MTNs' performance characteristics, which can be used for guiding the development of the next-generation MTNs.

In this paper, we investigate the performance of tactical radio networks, particularly for communication scenarios where multihop relaying along with spatial reuse techniques are applied. The tactical scenarios of concern have a diverse range of reliability and/or delay requirements. We first employ cross-layer protocol architecture with integrated time-division-multiple-access-based fast packet forwarding and automatic-repeat-request-based multihop error control to support tactical applications with a diverse range of reliability and/or delay requirements. Then, we develop analytical models to study the link-level and network-wide behaviors of the tactical radio networks. The developed models capture the effects of end-to-end channel memory due to multihop relays, interference due to spatial reuse, uneven interslot delays due to fast packet forwarding, and jamming attack in hostile network environment. It is shown that the protocol architecture with proposed fast packet forwarding and multihop error control mechanisms can significantly improve communication performance and support delay-sensitive applications. In addition, it is demonstrated that cross-layer adaptivity, where different network environments require different logical topologies and radio modes, is needed to achieve best performance tradeoffs among throughput, efficiency, delivery ratio, and transport capacity.