Joint Communication and Sensing (JCAS) represents a key paradigm shift for future wireless systems, enabling efficient use of hardware and spectral resources. However, the integration of these two functions creates a fundamental challenge in resource management, as communication and sensing have conflicting performance objectives. This thesis addresses this challenge by developing a practical, simulation-based framework to optimize resource allocation in a communication-centric, mmWave JCAS system. A comprehensive system-level model is developed, which integrates time-frequency allocation, gain sharing, and beamforming at the Physical Resource Block - Transmission Time Interval level. Using this model, extensive Monte Carlo simulations are performed to characterize the trade-off between communication throughput and sensing accuracy, evaluated using practical metrics such as Root Mean Square Error (RMSE) and throughput. The results reveal that allocating joint resources in the center of the time-frequency grid outperforms the edge-based allocations using practical periodogram estimators. The proposed optimization tool successfully identifies the optimal resource parameters that maximize system performance under specific constraints. Furthermore, the analysis of an interference-nulling receiver provides critical insights into the trade-offs between interference suppression, noise enhancement, and beam robustness. Ultimately, this work provides a valuable design tool and a set of clear strategies for configuring the operational parameters of practical JCAS systems.

The sixth generation (6G) of mobile networks promises transformative capabilities in terms of, amount others, higher data rates, lower latency and ubiquitous coverage, but achieving these goals sustainably poses significant challenges. A promising solution lies in Cell-Free massive Multiple-Input Multiple-Output (CF-mMIMO) networks, where a dense deployment of geographically distributed Access Points (APs), coordinated by one or more Central Processing Units (CPUs), cooperatively serve User Equipments (UEs). While CF-mMIMO networks offer improved spectral efficiency and more spatially uniform service, their dense deployments can result in high energy consumption. As MNOs typically design their networks such that Quality of Service (QoS) targets are met during peak hours, the network is left significantly over-dimensioned during off-peak hours. This thesis addresses this inefficiency by proposing a low-complexity, heuristic Sleep Mode Management (SMM) algorithm that reduces energy usage by switching off unneeded APs while maintaining acceptable QoS and without compromising coverage.

The proposed SMM algorithm supports multiple AP power states: active, light sleep and deep sleep. It incorporates realistic transition times between these states. Relying solely on practically available information such as long-term Channel State Information (CSI) and previously achieved data rates, the algorithm dynamically decides which APs can be temporarily put into a sleep mode. Importantly, it ensures population coverage is maintained and it preserves UE QoS based on a 10th UE throughput percentile target.

The proposed SMM algorithm is evaluated using a system-level simulator that models a realistic scenario based on the city center of Amsterdam, including lamppost-based AP deployments and a realistic basis for the spatial traffic distribution and daily traffic fluctuations. Simulation results demonstrate that the proposed SMM algorithm reduces the daily energy consumption of a CF-mMIMO network by up to 17.11\% with the best overall configuration. If the parameters of the SMM algorithm are allowed to be adaptively tuned to the traffic load, the daily energy consumption can be reduced by up to 21.54%. This thesis not only contributes a novel SMM algorithm but also provides guidance for AP deployment strategies in 6G CF-mMIMO networks, determining that a higher number of lower-antenna-count APs can yield better QoS than fewer, more higher-antenna-count APs at the cost of energy efficiency.

Reconfigurable Intelligent Surfaces (RIS) have emerged as a promising technology to enhance the performance of next-generation wireless networks, particularly in terms of increasing coverage and enhancing user throughput. However, effective radio resource management for RIS-aided mobile networks introduces significant complexity due to the high dimensionality and the interdependency of these tasks.

This thesis proposes a solution that leverages the machine learning-based framework of Graph Neural Networks (GNNs) to jointly optimize radio resource management tasks in a multi-user scenario.
The proposed model is capable of performing user scheduling, beamforming, and RIS configuration based on full channel information that maximizes proportional fairness. Additionally, the model also supports implicit power allocation.

A comprehensive simulation environment emulating a dense urban mobile network is developed to train and test the model. The performance across various mobile network deployments is evaluated and compared to an accurate existing method.
Results demonstrate that the proposed GNN-based solution manages to achieve a large fraction of the user throughput gain achieved using the existing method in significantly less computation time, showcasing its potential for real-time radio resource management in future 6G networks.

This thesis explores Radio Resource Management (RRM) techniques for Joint Communication and Sensing (JCAS) in beyond-5G/6G networks. JCAS integrates communication and sensing functionalities into a unified network, promising enhanced spectral efficiency, reduced system costs, and improved performance in various applications such as autonomous vehicles and smart cities. However, efficiently managing the dual requirements of communication and sensing in a shared network is a significant challenge, especially in dense, multi-cell environments.
The work presents a novel approach to JCAS by developing advanced resource management algorithms aimed at optimizing communication throughput and sensing accuracy, with a focus on target detection as the primary sensing task. The proposed algorithms encompass topology selection, dynamic node contribution management, and joint scheduling of sensing and communication tasks. Through extensive simulations, we analyze the trade-offs between communication and sensing performance, considering metrics such as the average and 10th percentile user throughput and probability of detection. The results demonstrate the effectiveness of the proposed strategies in managing interference and improving system performance in a cooperative multi-cell JCAS network.
This study contributes to the growing body of research on JCAS by addressing key limitations in existing works, such as the lack of cooperative sensing and the need for real-time dynamic resource allocation algorithms. The findings provide practical insights for network operators and lay the groundwork for future research into more complex JCAS applications and further optimization of resource management techniques.

Reconfigurable Intelligent Surfaces (RIS) are envisioned to become a pivotal transformative technology within the realm of 6G mobile networks. In this study, we introduce three heuristic algorithms designed to optimize radio resource management, ultimately enhancing throughput within a RIS-enhanced mobile network. Our findings demonstrate that the algorithm which holistically optimizes user scheduling, RIS-UE association, cell precoding matrices, and RIS configuration matrices outperforms alternative strategies.

Moreover, our investigation delves into uncovering the most effective applications of RIS. This involves a thorough performance comparison of the algorithms across diverse scenarios, encompassing varying RIS deployment configurations (position and orientation), number of users in a network, and number of RIS elements. Additionally, we model the influence of blocker loss—characterized by blocker presence probability and strength—on throughput performance.

In the wake of our study, it becomes evident that RIS exhibits the most promising potential in scenarios involving MU-MIMO configurations, whether within single-cell or multi-cell layouts, and for both indoor and outdoor user settings. However, for SU-MIMO cases, RIS-induced throughput enhancement manifests exclusively in single-cell layouts, and particularly benefits outdoor users in environments marked by substantial blocker strength.

Increasing wireless communication requirements of data rates, capacity and coverage, and evolution and maturation of wireless equipment prompt wireless communication research insight concentrating on millimeter wave (mm-wave) frequency. However, high reflection coefficients and high path loss cause large shadow areas (e.g. behind the buildings) and poor coverage, leading to constraints on wireless connectivity and effectiveness of wireless communication. Intelligent Reflecting Surface (IRS) is a revolutionizing technology in 6th-generation mobile networks (6G), which achieves extended coverage with reducing construction and electricity costs via its characteristics of passive beamforming and proper deployment, and auxiliary of Ray Tracing (RT) facilitates obtention and analysis of channel state information (CSI) at different locations. With the objective of developing a flexible RT tool and a novel methodology for optimal IRS deployment to maximize coverage in non-line-of-sight (NLOS) areas from the BS, a new RT simulation model is built in this thesis project in accordance with measured data, leading to improved reliability and accuracy of 5.3\% maximum error in path loss. And a novel, but preliminary, weight graph methodology is proposed for tackling the IRS deployment problem for coverage extension in NLOS areas quasi-optimally. To integrate the IRS with RT simulation, a first-time comparison between metal reflectors and IRS under realistic EM effects is exploited. The obtained simulation results unveil that deploying IRS with the proposed weight graph methodology facilitates wireless coverage improvement, and the coverage probability increased from 0% to 96.23% with a threshold of -75 dBm under 28 GHz in a selected Region of Interest (RoI).

Hybrid beamforming (HBF) architecture provides promising trade-offs between the system performance and computational/hardware complexity in practical implementations of millimeter wave (mm-Wave) massive multiple-input multiple-output (mMIMO) 5th generation (5G) mobile networks compared to its fully digital beamforming (DBF) counterpart. In this thesis, we investigate the future applicability of deploying hybrid beamforming architectures with subarray beam pattern shaping for mm-Wave 5G base stations in spatially heterogeneous user distributions and different propagation scenarios. We propose HBF structures with a cosecant-squared pattern and a flat-top pattern as well as their HBF and DBF benchmarks. In addition to the uniform user distribution, three non-uniform user distributions, i.e., the near-site distribution, the cell-center distribution, and the cell-edge distribution are proposed to represent the traffic flow and mobility of users due to festivals and holidays. We evaluate the performance in a novel 5G new radio (NR) system-level simulation (SLS) model. Numerical results show that the HBF architecture with a cosecant-squared subarray beam pattern is more robust against differences in spatially heterogeneous traffics than the flat-top HBF and benchmark HBF under the line-of-sight (LoS) propagation scenario. Under the non-line-of-sight (NLoS) propagation scenario, more deterministic environment information and radio channel modeling are required to improve the system performance of the shaped HBF beamforming technique.

The Fifth Generation (5G) of mobile networks exploit both sub-6 GHz and millimeter-wave (mmWave) spectrum. The sub-6 GHz spectrum comprises frequencies up to 6 GHz and provides large geographical coverage for radio signal in 5G. The mm-wave spectrum on the other hand, comprises higher frequencies ranging from 24 GHz -100 GHz and plays a major role in serving high data rates for the 5G technology. The evolution of 5G plays a pivotal role in the realization of challenging applications like Social XR conferences, which requires the network to deal with heavy traffic while maintaining low end-to-end latencies. The right configuration of the radio access network becomes crucial for such applications. The introduction of Massive Multiple Input Multiple Output (MIMO) technology in the radio network, concentrates the signal energy to the target user, which significantly improves the throughput and efficiency of the system. The quality and capacity of the radio channels also depend on the downlink Channel State Information (CSI). The CSI when obtained accurately at the Base Station (BS), plays a significant role in reaping the best benets out of Multiple Input Multiple Output (MIMO) technology. This thesis explores (or assesses) the different options and configurations of CSI feedback using an indoor Social Extended Reality (XR) conference application scenario. The performance analysis of the radio downlink while using the latest 5G New Radio (NR) Types I and II CSI which use a DFT-based codebook are detailed. The impact of the codebook-related configurable parameter of Rotation Factor (RF), the performance variations while using a `fixed-RF' for all the UEs compared to the more flexible `adaptive-RF', different beamforming technologies (Single-User MIMO (SU-MIMO) and Multi-User MIMO (MU-MIMO)), transmission ranks and co-scheduling parameter values are assessed using the key performance metric of Packet Loss Ratio (PLR). The frequency bands of 3.5 GHz (sub-6 GHz spectrum) and 26 GHz (mmWave spectrum) are chosen for the thesis and performance variations between the two bands are studied. The key insight from the thesis research is that the `adaptive-RF' case gives the optimal performance for the considered Social XR scenario when we set the right co-scheduling parameters (which balance the encountered interference and frequency of co-scheduling).

As a result of a global pandemic, there has been an increasing interest in tools for remote video conferencing and collaboration. One of these new innovations is social eXtended Reality (XR). By combining Virtual Reality (VR) and Augmented Reality (AR) technologies, social XR can provide a more immersive experience than any other VR application by giving users at different locations the chance to virtually gather in real-time. But such applications impose enormous requirements on computational and communication resources. 5th Generation (5G) mobile networks are targeted as solution to provide ultra-low latency and ultra-high throughput for social XR. In current research, many optimisations are aimed at VR applications such as on-demand streaming, while there is a lack of solutions for real-time user-interactive applications like social XR. In this graduation project we develop and assess cross-layer solutions for optimised scheduling of social XR applications in 5G networks. An existing framework for simulating social XR conference applications serves as the basis for our modelling approach. We devise different schedulers, that utilise cross-layer information in the form of the video frametype and frame-level End-to-End (E2E) latency budgets rather than packet-level latency budgets purely within the Radio Access Network (RAN). In contrast to previous work, we create the VR traffic based on real video data and develop tools to model the packet dispersion caused by multi-hop transmission over the internet towards the RAN. We study the effect of various system and traffic parameters on the Quality of Service (QoS) and perceived Quality of Experience (QoE) in the context of social XR applications through an extensive sensitivity analysis. Herein we also assess the performance impact of different types of cross-layer packet schedulers. Further, we gain insights into the correlation between the network QoS and perceived QoE by end users which are the key in future cross-layer implementations for social XR.

The Fifth Generation (5G) network is expected to support three main service categories namely enhanced Mobile BroadBand (eMBB), Ultra-Reliable and Low-Latency Communications (URLLC) and massive Machine Type Communications (mMTC), where each service group has a different Quality of Service (QoS) requirements i.e. the eMBB service group has a high throughput requirement, the URLLC service group require very low-latency and highly reliable transmissions and the mMTC service group do not have a strict performance requirement but have massive connection of devices in the network.

5G enables many vertical domains with these three service categories, such as, smart cities. In a smart city environment, there are applications from all three service categories such as massive connectivity of the sensors for waste managements or monitoring environmental conditions, video surveillance along the city streets and many more. This study addresses the problem of managing applications from the three service categories on the same physical network infrastructure at the Radio Access Network (RAN). Two different scenarios, one with and one without an emergency incident, are considered to find the impact of the incident in the network.

In 5G networks, many new features are introduced such as, flexible numerology, mini-slot based scheduling, BandWidth Parts (BWPs) and RAN slicing. The key objective of this study is to assess the 5G RAN features in terms of achieving the performance requirements of the considered applications, simultaneously. To do so, different RAN configurations are modelled where, a RAN configuration consists of one or multiple RAN features. The evaluation is done by simulating the different possible RAN configurations. The simulations are performed using an existing 5G system-level simulator which is substantially upgraded with the 5G RAN features and is modified to the considered smart city urban macro-cellular environment and with the considered traffic models for each considered application.

To evaluate the performance of each considered application, different performance metrics are defined based on the application requirements. The benefits and/or losses of different RAN features are found and then different RAN configurations are considered with the combinations of RAN features based on the evaluation of each RAN feature. For all different RAN configurations with combination of features, the performance metrics are evaluated and compared with each other to determine the best-performing configurations for the smart city environment, for the scenarios with and without an incident.

In this thesis, we design and assess a multi-slice resource allocation framework that is based on machine learning techniques (subset of artificial intelligence techniques). The proposed framework employs two machine learning techniques namely, artificial neural networks and reinforcement learning for resource management in sliced RAN. Alternative multi-slice resource allocation methods that involve only artificial neural networks but not reinforcement learning are also defined.

5G networks are expected to be used in many markets, one of which is the Factories of the Future (FoF). In the FoF, applications like regular monitoring and controlling of components (e.g. temperature) are introducing the need of massive deployment of sensors and actuators. Additionally, sensors which are monitoring time-critical components of the factory (e.g. pressure in power plants) should experience low delays with a high reliability. The current LTE-based technology for machine-type communications, namely Cat-M1, imposes limitations in supporting massive connectivity and application with low delay and high reliability requirements, primarily due to the so-called ‘Random Access’ (RA) procedure which is used by the devices to establish a connection with the base station, before the actual transmission of their data. The key objectives of this study are to assess the RA procedure currently used in Cat-M1 networks as a baseline, and furthermore design and evaluate a new RA procedure for 5G networks, which targets typical FoF applications and their requirements. Based on the simulation assessment of the procedures studied, it was found that the so-called ‘Two-Step RA procedure’ is performing the best regarding end-to-end delay. Specifically, it was found that e.g. in a FoF network with 3000 devices an end-to-end delay of 16 ms can be achieved with 99.99% reliability, introducing a gain of 64 ms compared to the end-to-end delay in Cat-M1 networks.

According to its creators, ICN is designed to fit the way we use the internet better than IP currently does.
The use of named data and distributed network layer caching may provide a more efficient utilization of
network resources due to the stateful forwarding plane which allows data to be retrieved from caches
close by the requester, while also providing a higher content delivery performance in terms of content
retrieval delay. Since the IoT is expected to connect billions of devices to the internet, a resource-efficient
network paradigm is needed to cope with the corresponding enormous traffic increase. IoT
deployments also typically follow a distributed data generation and retrieval paradigm, which could
benefit from ICN’s in-network caching approach and stateful forwarding logic. This thesis focuses on
assessing whether ICN is advantageous for the IoT in these aspects, by comparing an ICN approach
to an IP approach for IoT applications.

5G, the next generation of mobile network, is expected to be launched commercially around 2020. Compared to the present generation – 4G mobile network, a significant improvement in terms of performance and reliability is considered for 5G. One of the important factor in the design of 5G is – about 10 times lower packet latency than 4G. Some of the use cases identified for 5G require packet latency as low as 1 ms. Such stringent latency targets are essential to enable new services like virtual reality streaming of live content over mobile network, automated vehicle platooning over mobile network and tactile internet where machines and tools can be controlled remotely with extreme responsiveness over the mobile network.
The main goal of this thesis is to understand how packet latency is affected by the various factors observed in a realistic environment. In contrast to lab environments, where the packet latency reported would be very low, a consolidated study on the various factors affecting packet latency in a 4G (LTE) network in a realistic environment is missing. To this extent, the results of this work have enabled to identify the various factors affecting packet latency in a realistic 4G network. This further led to identifying the latency contribution of the various components to the overall packet latency. Later on, two different latency reduction techniques were evaluated to verify the possible latency reduction achievable on a 4G network, using those two techniques.
To reduce packet latency to achieve the latency targets for 5G, first it was necessary to identify how packet latency is caused and affected in a 4G network. This work was aimed at achieving this goal. As the latency reduction techniques were evaluated at their best configuration in terms of latency, results from the latency reduction techniques also identifies the lower limit of latency improvement achievable in a 4G network. The inference from the results suggests that in order to achieve the latency targets specified for 5G networks, a redesigned radio access technology of 4G is essential.

The environmental policy of KPN has an objective of reducing CO2 emissions and energy consumption to contribute in making the planet more sustainable. This thesis project researches the potential for reducing energy consumption and, consequently, the operational costs of the KPN LTE network. There are various resources in a mobile network which consume energy and switching off these resources will lead to a reduced energy consumption. The trade-off in switching off these resources involves a potential impact on the capacity and the quality (throughput) offered by the mobile network. In this thesis we quantify the attainable savings in operational cost and/or energy consumption while still satisfying the quality of service requirements set by KPN. Five different operational cost saving schemes are analyzed on three different timescales in this research. The results of this research indicate that there is ample opportunity present for KPN to reduce their operational cost and/or energy consumption with minimal impact on the quality of service. The results further indicate that maximum savings are attainable in the live KPN network by effectuating the ’turning off carriers’ scheme, in which the capacity carrier is turned off at the base stations. The reduction in energy consumption and operational cost on effectuating this energy reduction scheme is in the range of 21-70% with respect to the reference scenario for each of the timescale. Furthermore, the results also indicate that of the five operational cost saving schemes investigated, the ‘turning off carriers’ scheme has the lowest impact on the experienced quality of service.

Narrowband Internet-of-Things (NB-IoT), recently introduced by 3GPP, is a relevant Radio Access Technology (RAT) solution for deployment within smart grids, the electricity grids of the future, due to the need to provide low-cost connectivity to a large number of smart meters installed in households. Outage Restoration and Management (ORM), a smart grid use case, involves the smart meter User Equipment (UE) sending a notification message to the utility upon the detection of a loss or restoration of power. ORM is an effective way for utilities to quickly detect, localise and restore a power outage. However, depending on the extent of the power outage, a near simultaneous network access by multiple UEs may occur, leading to resource congestion, particularly of the so-called ‘random access channel’. This may impact the reliability, i.e. the percentage of total notifications successfully delivered within a certain transfer delay target, and in turn, the accuracy of the power outage localisation. Consequently, the maximisation of the ORM reliability performance for a technology like NB-IoT becomes a challenge, given that such use cases, though relevant, were not considered in its design phase.
The main goal of this thesis is the optimisation of the NB-IoT network configuration, with a focus on packet scheduling, in order to maximise the ORM reliability performance. To this extent, a system-level simulation model is developed and implemented, incorporating realistic characteristics of energy distribution and mobile networks in four different environments (rural, suburban, urban and dense urban), the traffic characteristics of ORM and the relevant 3GPP specifications of NB-IoT. Additionally, a set of candidate time-frequency domain packet schedulers are proposed. A sensitivity analysis of key network configuration components is performed for a set of power outage scenarios i.e. network loads, the associated optimisation trade-offs are highlighted and a near-optimal network configuration is derived.
Based on the sensitivity analysis, a proposed scheduler which prioritises UEs based on a combination of the Earliest Due Date First (EDDF) and Shortest Processing Time First (SPTF) principles, and assigns each UE a single uplink subcarrier with a subcarrier spacing of 3.75 kHz, performs best amongst all the candidate schedulers. Furthermore, the achieved reliability performance is close to 100% for all the considered power outage scenarios in the rural and dense urban environments. In the suburban and urban environments, close to 100% reliability is achieved for the majority of the power outage scenarios.

This thesis project researches the effect of the optimisation time interval on the performance of a self-optimised mobile network. The goal of the thesis is to ascertain if there exists an optimal time interval for the self-optimisation of the KPN network, and what that interval is. In order to research this question, the project uses data from the KPN network as input, and sets up a simulation study in MATLAB. Two areas in the Netherlands are considered in this study – Friesland and Purmerend. The self-optimisation of the network is carried out through the modification of three optimisation parameters – antenna tilt, RS power, and Cell Individual Offset. The scope of the study is limited to LTE in the downlink, for the 800 MHz band. The bandwidth used in this study is 10 MHz. The performance of the mobile network has been studied using KPIs such as 10th throughput percentile, coverage failure rate, call drop rate, and load. In the end, the study analyses the results for each area, for the self-optimisation carried out by modifying the three parameters over several different optimisation time intervals, and discusses their impact on the performance of the network. A comparison has also been drawn between the performance of a self-optimised network and an un-optimised network, to highlight the gains achieved with SON. Finally, recommendations are made regarding a suitable time interval, and a relative comparison between suitability of the three optimisation parameters has been drawn.
The study finds that a suitable time interval for optimisation does exist, and is 240 minutes, for both the simulation areas. The study finds RS power to be the most suitable parameter for self-optimisation, in both the areas. However, the research runs into some unexpected results with respect to the optimisations using tilt angle, and has been discussed in detail in the report. Significant gains are observed with SON, as compared to the case of ‘No SON’ or an un-optimised network.