The growing demand for ubiquitous broadband network connectivity and continuously falling prices in hardware operating on the unlicensed bands have put Wi-Fi technology in a position to lead the way in rapid innovation towards high performance wireless for the future. The success story of Wi-Fi contributed to the development of widespread variety of options for unlicensed access (e.g., Bluetooth, Zigbee) and has even sparked regulatory bodies in several countries to permit access to unlicensed devices in portions of the spectrum initially licensed to TV services. In this thesis we present novel spectrum management algorithms for networks employing 802.11 and TV white spaces broadly aimed at efficient use of spectrum under consideration, lower contention (interference) and high performance. One of the target scenarios of this thesis is neighbourhood or citywide wireless access. For this, we propose the use of IEEE 802.11-based multi-radio wireless mesh network using omnidirectional antennae. We develop a novel scalable protocol termed LCAP for efficient and adaptive distributed multi-radio channel allocation. In LCAP, nodes autonomously learn their channel allocation based on neighbourhood and channel usage information. This information is obtained via a novel neighbour discovery protocol, which is effective even when nodes do not share a common channel. Extensive simulation-based evaluation of LCAP relative to the state-of-the-art Asynchronous Distributed Colouring (ADC) protocol demonstrates that LCAP is able to achieve its stated objectives. These objectives include efficient channel utilisation across diverse traffic patterns, protocol scalability and adaptivity to factors such as external interference. Motivated by the non-stationary nature of the network scenario and the resulting difficulty of establishing convergence of LCAP, we consider a deterministic alternative. This approach employs a novel distributed priority-based mechanism where nodes decide on their channel allocations based on only local information. Key enabler of this approach is our neighbour discovery mechanism. We show via simulations that this mechanism exhibits similar performance to LCAP. Another application scenario considered in this thesis is broadband access to rural areas. For such scenarios, we consider the use of long-distance 802.11 mesh networks and present a novel mechanism to address the channel allocation problem in a traffic-aware manner. The proposed approach employs a multi-radio architecture using directional antennae. Under this architecture, we exploit the capability of the 802.11 hardware to use different channel widths and assign widths to links based on their relative traffic volume such that side-lobe interference is mitigated. We show that this problem is NP-complete and propose a polynomial time, greedy channel allocation algorithm that guarantees valid channel allocations for each node. Evaluation of the proposed algorithm via simulations of real network topologies shows that it consistently outperforms fixed width allocation due to its ability to adapt to spatio-temporal variations in traffic demands. Finally, we consider the use of TV-white-spaces to increase throughput for in-home wireless networking and relieve the already congested unlicensed bands. To the best of our knowledge, our work is the first to develop a scalable micro auctioning mechanism for sharing of TV white space spectrum through a geolocation database. The goal of our approach is to minimise contention among secondary users, while not interfering with primary users of TV white space spectrum (TV receivers and microphone users). It enables interference-free and dynamic sharing of TVWS among home networks with heterogeneous spectrum demands, while resulting in revenue generation for database and broadband providers. Using white space availability maps from the UK, we validate our approach in real rural, urban and dense-urban residential scenarios. Our results show that our mechanism is able to achieve its stated objectives of attractiveness to both the database provider and spectrum requesters, scalability and efficiency for dynamic spectrum distribution in an interference-free manner.

The advent of 6th Generation (6G) wireless systems and the increasing demand for spectrum to accommodate a growing number of users and diverse services have necessitated novel ap- proaches to spectrum sharing. Among these approaches, distributed spectrum sharing offers the most flexibility by allowing real-time spectrum use based on user demand and network con- straints. However, this approach presents significant challenges due to the probabilistic nature of system dynamics and the autonomous behavior of each incumbent, which require advanced strategies to predict and manage spectrum usage effectively. Listen-Before-Talk (LBT) is the most widely adopted method for distributed spectrum sharing in unlicensed bands. While LBT has been extensively studied in the context of Wireless Fidelity (Wi-Fi), providing key insights into its performance under various conditions, its application in synchronized, slot-scheduled sys- tems like New Radio (NR) Unlicensed (NR-U) remains underexplored. This gap exists primarily due to the lack of hardware testbeds and system-level simulation platforms that are essential for evaluating the effectiveness of LBT in NR-U and for developing improved methods for operating in shared spectrums with deterministic worst-case delays. This thesis addresses the existing gap by proposing a reference architecture for spectrum sharing based on 5th Generation (5G) NR-U to facilitate further research and experimentation in distributed spectrum sharing. The approach taken in this thesis is threefold: (i) the establishment of a system architecture for an end-to-end 5G NR-U system based on existing work in hardware and simulation models; (ii) the realization of this system model on the Network Simulator 3 (ns-3) discrete-event simulator by leveraging developments from the 5G Long-Term Evolution (LTE) Enhanced Packet Core (EPC) Network Simulator (LENA) (5G-LENA) system architecture; and (iii) the conceptual design for implement- ing the Physical (PHY) layer of a 5G NR-U system using Software-Defined Radios (SDRs) and the OpenAirInterface (OAI) 5G software platform. A key novelty of this reference architecture is the proposed mitigation of LBT latency in split architectures with SDRs and General-Purpose Processors (GPPs). The LBT block is designed for implementation within the Field Program- ming Gate Array (FPGA) of Universal Software Radio Peripheral (USRP) SDRs, thereby enabling heterogeneous coexistence experimentation with Common Off-the-Shelf (COTS) Wi-Fi Access Points (APs). The thesis presents a simulation-based experiment that optimizes traffic manage- ment to improve the ability to serve delay-critical traffic in NR-U systems operating under ho- mogeneous coexistence conditions. The thesis then outlines a reference design for exploring heterogeneous coexistence between Wi-Fi and NR-U in the sub-7 GHz spectrum. This concep- tual framework leverages a proposed hardware experimentation platform with SDRs. The in- frastructure supporting these simulations and proposed hardware experiments is envisioned as virtualized resources over the Commonwealth Cyber Initiative (CCI) xG Testbed, with potential extensions for advanced spectrum sharing use cases across indoor and outdoor testbed sites. The thesis outlines potential enhancements to this testbed, specifically toward spectrum sharing with scheduled-access systems. / Master of Science / As wireless communication demand grows with the development of 6G, finding efficient ways to share the limited available spectrum has become increasingly important. One promising ap- proach is distributed spectrum sharing, which allows dynamic use of the spectrum based on real-time demands. However, this method faces challenges due to the unpredictable behavior of different users and devices, requiring sophisticated strategies to manage spectrum usage effec- tively. Currently, the most common method for distributed spectrum sharing is LBT, widely used in Wi-Fi networks. Although LBT has been well-studied in these environments, its use in systems like NR-U – a variant of 5G designed for unlicensed spectrum—has not been thoroughly explored. This gap exists mainly because there are few hardware testbeds and simulation platforms avail- able to study how LBT and other methods might work in real-world systems. This thesis aims to address this gap by developing a standardized platform for testing and experimenting with 5G NR-U technologies. The work involves three key steps: (i) designing a comprehensive system architecture for 5G NR-U; (ii) implementing this system in a simulation environment to study its performance; and (iii) proposing a design for key components using SDR and open-source soft- ware, creating a foundation for future hardware-based testing. To demonstrate the capabilities of this new platform, we conducted a simulation-based experiment focused on optimizing traffic management in NR-U systems to better handle delay-sensitive communications. Although no hardware experiments were conducted, the thesis provides a conceptual framework for future studies exploring how Wi-Fi and NR-U could coexist in the same frequency bands using the pro- posed hardware platform. The thesis concludes with suggestions for future improvements to the testbed, particularly in advancing spectrum sharing techniques with scheduled-access systems.