Quality-of-service aware resource allocation for integrated fiber-wireless (FiWi) access networks

Abstract

This thesis aims to propose an efficient resource allocation framework to ensure the quality of service (QoS) to the users. The frameworks proposed in this dissertation can be efficiently utilized to provide QoS guarantees by efficient resource allocation not only in the static FiWi network scenarios but also for the mobile vehicular networks. Specifically, this thesis is structured into four sections, namely energy resource allocation for green FiWi networks, QoS-aware high throughput FiWi network, latency reduction in FiWi-based vehicular networks, and emergency path planning for safe-driving in transportation Metaverse (TransVerse). The first section of this dissertation analyzes the techno, socio, and economic impact of energy resource allocation and provides 24/7 services to the users. In order to do so, we utilize a 3rd generation partnership project (3GPP) based model to generate the user request at the optical network unit (ONU) and access point (AP) collectively known as ONU-AP. These requests were used to calculate the load of the ONU-AP and further model the energy requirement, as well as allocate energy resources to provide 24/7 services to the users. In the second section, we extend the above work by utilizing the last 10 years’ data on available solar power to predict the amount of solar power that can be generated on a particular day. This would reduce the amount of charging and discharging cycles that are required by the battery and would improve the lifetime of the batteries. In a communication network, maintaining the latency and reliability requirements along with throughput is also crucial. Consequently, the second section of the thesis utilizes key performance indicators (KPIs) such as latency, reliability, and data rate-based classification to classify the data track at the APs into different services, such as voice, video, and best track. Based on this classification, the data traffic is assigned to different traffic containers (T-CONTs) of the PON. Depending on the KPIs, a priority-based bandwidth allocation scheme is designed to improve the QoS of the users. Moreover, during high mobility scenarios, meeting the QoS requirements becomes all the more di cult. Thus, the later part of this thesis extends the QoS guarantees for connected vehicular networks. Consequently, an integrated next-generation passive optical network 2 (NG-PON2) and IEEE 802.11p-based vehicle-to-infrastructure (V2I) network is proposed to facilitate the stringent requirements of sixth-generation (6G) vehicular networks. This work minimizes the latency of the FiWi-based V2I network by utilising a machine learning (ML) based T-CONT priority assignment wavelength allocation algorithm that minimizes the number of wavelength switching instances in the PON, sub- sequentially reducing the latency of the network. We further improve the performance of vehicular network by utilizing vehicle-to-vehicle (V2V) services along with V2I services. This requires the selection of an optimal cluster head for the offloading of V2V traffic. In order to do so, we utilize a double deep Q network (DDQN)-based cluster head selection policy to select the cluster head to minimize the energy and latency of the vehicular network. Furthermore, as a part of the 6G network, the research community has been recently focusing on emerging technologies such as digital twins (DTs), blockchain, extended reality, etc. Thus, in the last section, we exploit a transportation Metaverse (TransVerse)-enabled vehicular network to offload real-world vehicular data tra c to the vehicular cloud network and allocate physical and virtual resources in the network. We consider a pre-emptive emergency situation and formulate a utility-to-cost ratio optimization problem to minimize the collision risk along with the latency and computation cost of the network. Further, we utilize the prior information processing in the DT-enabled vehicular traffic to enhance the lane-change experience of the users. The frameworks proposed in this dissertation can be efficiently utilized to provide QoS guarantees by efficient resource allocation not only in the static FiWi network scenarios but also for the mobile vehicular networks. This will be helpful for a communication engineer to design a sustainable system that exploits new technologies, such as DT-enabled Metaverse, to guarantee the required QoS to the end users.