<p> Wireless metropolitan area networking based on IEEE 802.16 is expected to be widely used for creating wide-area wireless backhaul networks, where each subscriber station (SS) is responsible for forwarding traffic for a number of connections. Quality of Service (QoS) provisioning is an important aspect in such networks. The IEEE 802.16 standard specifies that the bandwidth requests sent by the SS are for individual connections and pass only the number of bytes requested from each connection. This is inefficient for backhaul networks where each SS may be responsible for forwarding packets for a relatively large number of connections and the bandwidth request messages consume much bandwidth unnecessarily. Furthermore, the standard does not include latency information, which makes it difficult for the base station (BS) to schedule real-time traffic. </p> <p> In this thesis we study real-time voice traffic support in IEEE 802.16-based backhaul networks. We propose a simple enhancement to the bandwidth request mechanism in 802.16 for supporting packet voice traffic. First, the SS combines the bandwidth requests of multiple voice connections, which are associated to it and have the same traffic parameters, and aggregates the bandwidth requests to the BS. This makes the bandwidth request process more efficient by saving transmission time of both the BS and the SSs. Second, in order to facilitate the BS to make resource allocation decisions, the aggregate bandwidth requests include information about the latency requirements of buffered real-time packets at the SSs. We propose three different bandwidth request and packet scheduling schemes, each of which requires a different amount of information in the bandwidth requests. Our results show that the proposed bandwidth request and scheduling schemes achieve significantly lower packet loss probability than standard 802.16 bandwidth requests and weighted round robin. The results further show that there is an optimum point about how much delay information the SS should report to the BS in order to best utilize the uplink resources while providing satisfactory real-time performance for the voice traffic. </p> / Thesis / Master of Applied Science (MASc)

Yes / Non-terrestrial Network (NTN), utilizing highaltitude platforms (HAP)-based free-space optical (FSO) backhaul and unmanned aerial vehicles (UAV) for last-mile access, is a feasible and promising architecture to achieve high data rate and seamless network coverage in the future 6G era. Effective resource allocation emerges as a pivotal concern for such networks. This paper addresses the data allocation issue for FSO backhaul from the HAP to multiple UAV-mounted base stations (BSs) under the constraints of ground users’ requested data rates. We introduce frame allocation schemes (FAS), including rate adaptation with constraints (RAC)- and rate/power adaptation (RPA)-aided FAS. The key idea of these schemes is to allocate data frames effectively based on UAV’s turbulence channel conditions, which aims to (i) guarantee the quality of services (QoS), (ii) retain both latency and throughput fairness, and (iii) minimize the transmitted power. Furthermore, the performance of these schemes is also analyzed under the impact of imperfect channel state information (CSI). We newly derive the channel probability density function (PDF) and the cumulative density function (CDF), considering the imperfect CSI due to channel estimation and quantization errors. Capitalizing on the derived PDF and CDF, different performance metrics are analytically obtained, incorporating combined effects of cloud coverage, transceiver misalignment, Fisher-Snedecor F turbulence, and angle-of-arrival (AoA) fluctuations. Numerical results demonstrate the effectiveness of our design proposals over the state-of-the-art. Finally, Monte Carlo simulations are employed to validate the analysis.