Doctorate thesis defense of Amal Sellami

Doctorate thesis defense on May 26^th 2025 at 09H30 AM ,in Amphitheater Ibn Khaldoun, SUP'COM 2.

Entitled : Localization for 5G and IoT Wireless Systems

Presented by : Amal Sellam

Committee

Abstract

Localization is a key component of the 5th Generation (5G) systems because of its obvious potential for communication performance enhancement. The advanced wireless technologies featured for 5G such as network densification, the use of millimeter waves (mmWaves), massive Multiple Input Multiple Output (MIMO) deployment that provide a high angular resolution, high bandwidth, cooperation and Device to Device (D2D) links, offer promising capabilities for accuracy improvement.

5G networks support heterogeneous users as they serve both low-power and high-capacity devices. We therefore consider both subsystem requirements and exploit such heterogeneity to enhance localization performance. Furthermore, the expanding context-aware services, such as Intelligent Transportation (IT), Unmanned Aerial Vehicles (UAVs), industrial, and other Internet of Things (IoT) applications, need location awareness.

In this thesis, we contribute to the design and implementation of massive MIMO 5G localization methods by focusing on challenges associated with this system. Several contributions are proposed which combine the received signal power, for distance recovery, and direction of arrival at the antenna, in order to reduce ambiguity in the User Equipment (UE) position while reducing the search space and then computational complexity.

In the first proposed scheme, we consider a conventional cellular system and propose a multi-stage localization method that enables reducing complexity by narrowing the search space and adapting the sub-antenna size of each stage. This approach shows strong robustness to channel uncertainties.

In the second contribution, we address the challenge of outdoor localization with a particular focus on scenarios where the gNB faces harsh channel conditions from the UE. We propose two neighbor-aided approaches and leverage the capabilities of neighbor discovery and oriented beamforming to enhance the accuracy of UE positioning. The methods begin with a neighbor discovery step, where reference signal power measurements allow distance estimation between the T-UE and its neighboring anchors. These distances are then combined in two different ways with optimized beamforming to estimate the UE location.

In the last contribution, to facilitate seamless communication in areas where traditional infrastructure is not feasible or cost-effective, UAVs are introduced as aerial gNBs to offer a more flexible way to improve signal coverage and thus enhance location accuracy. A strategy involving a single gNB paired with a UAV at two positions is considered. A fusion of trilateration and beamforming then crystallizes the UE's final position. A collaborative algorithm is proposed that leverages Reference Signal Strength (RSS) measurements provided by both a Line of Sight (LoS) UAV and a gNB to accurately locate a static UE. Distance estimation is associated with trilateration and beamforming to locate the UE. The method demonstrates enhanced sub-meter accuracy performance when compared to previous methods. The analysis also reveals certain areas of potential improvement, particularly in high SNR ranges.

In summary, we show throughout this thesis that 5G mmWaves systems can accurately localize a UE with a sub-meter accuracy, which plays an important role in the development of future applications and services. This level of accuracy enhances the deployment of autonomous vehicles and industrial automation systems that require accurate localization for safety and operation. Ultimately, this result plays a central role in communication network optimization, unlocking opportunities that were not available in previous generations, thus facilitating a more connected user experience.

Keywords

Localization, 5G, IoT, Massive MIMO, mmWaves, Neighbor Discovery, UAVs