Conference Publications

Permanent URI for this collection

http://hdl.handle.net/10204/11486

Browse

Now showing 1 - 20 of 187

4G RAN infrastructure sharing by 5G virtualized mobile network operators: A tutorial
(2021-09) Mamushiane, Lusani; Mboweni, Lawrence S; Kobo, Hlabishi; Mudumbe, Mduduzi J; Mwangama, J; Lysko, Albert A
Active radio access network (RAN) infrastructure sharing has emerged as a promising solution for efficient spectrum utilization, capital and operational cost savings, improved MVNO penetration rates and lower broadband retail prices in both emerging and developed markets. This paper presents a tutorial on the testbed implementation of an active RAN sharing architecture, leveraging multi-vendor virtualized 5G and 4G core networks running on commodity hardware and proprietary 4G RAN equipment (eNodeB). Troubleshooting techniques used for different implementation challenges encountered are also presented in this contribution. The performance of the proposed architecture was validated using end-user quality of experience (QoE) as the key performance indicator. The results show no performance degradation when RAN sharing is being utilized.
5G network slice resource overbooking: An opportunity for telcos to boost their revenue
(2022-08) Mamushiane, Lusani; Mwangama, J; Lysko, Albert A; Kobo, Hlabishi I
The global impact of COVID-19 has been unprecedented, with over-the-top (OTT) services consumption growing at a staggering rate. While OTT does consume revenue-generating data, OTT services are gradually substituting the traditional primary sources of revenue, voice and SMS services, with “freemium-based” alternatives such as WhatsApp and Telegram. This has driven telcos to reconsider their strategies and revenue sources. We believe that 5G network slicing (a type of 5G infrastructure sharing) is a potential solution that telcos could adopt to boost their revenue. In particular, this paper introduces the concept of network slice resource overbooking. We consider leveraging machine learning (ML) to maximise network resource utilisation which translates to maximum revenue gains for telcos. The concept of overbooking is unique and novel in network slicing. To realise this objective, we intend to build a mathematical model of the overbooking strategy, and integrate the model into a resource orchestration platform for evaluation on an emulated 3GPP (Release 16) compliant 5G testbed.
A brief performance comparison of bare-metal and kubernetes deployments for 5G Core Control plane network functions using Open5GS
(2024-10) Mukute, T; Santos de Brito, M; Lysko, Albert A; Mwangama, J; Magedanz, T
This paper investigates the performance difference of critical 5G User Equipment (UE) procedures when deployed on a Kubernetes platform versus a traditional bare-metal deployment. We leverage Open5GS, an open-source implementation of the 5G Core (5GC), to evaluate the impact of containerisation on key performance metrics. The research answers (i) how the performance of critical 5G UE procedures differs when 5GC is deployed on a Kubernetes environment compared to a traditional bare-metal deployment and (ii) provides the measurable cost introduced by Kubernetes in terms of key 5G performance metrics. We evaluated throughput and latency. The paper analyses the observed performance differences against theoretical expectations arising from the Kubernetes architecture overhead and insights from related work. Our study reveals a 7% performance degradation in throughput for UE procedures running on Kubernetes compared to bare-metal when handling more than 300 initiated UE devices.
A review of AI/ML algorithms for security enhancement in cloud computing with emphasis on artificial neural networks
(2024-11) Rakgoale, DM; Kobo, Hlabishi I; Mapundu, ZZ; Khosa, TN
Security in cloud computing is becoming increasingly important due to the scale and sensitivity of data handled. Artificial Intelligence and Machine Learning algorithms offer robust solutions for enhancing security measures in cloud environments. Traditional security methods often fail to safeguard private information and user anonymity against the ever-evolving cyber threats. As new attack methods and techniques continuously emerge, traditional approaches become inadequate. Embracing modern strategies can enhance resilience and better protect sensitive data. Utilising AI and machine learning for threat detection, automated responses, and predictive analytics can significantly improve contemporary security measures. The use of Artificial Intelligence and Machine Learning is essential in the current era of big data to handle and analyse enormous amount of cloud-based data quickly and accurately. In addition to the security challenges posed by cloud computing and Internet of Things (IoT) devices, the utilisation of AI by hackers remains an ongoing threat in the realm of cybersecurity. This paper review various Artificial Intelligence and Machine Learning algorithms, with a particular focus on the application of Artificial Neural Networks (ANNs). It further provides an analytical review of how ANNbased approaches contribute to an improvement of threat detection, anomaly detection in cloud computing, highlighting their Zamikhaya Z, Mapundu Tshwane University of Technology Pretoria,South Africa MapunduZ@tut.ac.za Artificial Intelligence (AI) enabled technologies empower security systems to identify patterns, anomalies, and potential threats across large datasets. By leveraging Machine Learning (ML) algorithms that analyse historical attack data, these systems can forecast future threats and enhance their defensive strategies proactively [1]. This capability not only improves the detection and response time to security incidents, but also enable organisations to adapt swiftly to evolving cyber threats. Moreover, AI enables security solutions to automate routine tasks, allowing human analysts to focus more on complex and strategic aspects of cybersecurity management [3]. Thus, integrating AI into security frameworks not only bolsters protection but also enhances overall operational efficiency and resilience against cyber threats. Furthermore, the convergence of Machine Learning and AI in cloud frameworks is expected to accelerate digital transformation and bring in an exciting period of efficiency and innovation [4]. effectiveness, potential challenges. Moreover, the advantages of artificial neural networks (ANNs) are discussed along with the current challenges encountered when applying these advanced models in cloud computing security.
A review of dynamic RRA techniques on 5G and beyond mobile networks
(2024-07) Nokane, Boikobo; Isong, B; Masonta, Moshe T
The 5G mobile network aims to enhance wireless communication by providing faster and more reliable connectivity. Open radio access network (RAN) architecture, which offers flexibility and innovation in Radio Resource Allocation (RRA), is central for optimal network performance. However, traditional RRA methods fall short of meeting the complex demands of 5G due to scalability issues. Incorporating machine learning (ML) techniques into open RAN can enhance adaptability and intelligence, ensuring that 5G networks meet high performance and service quality standards. This paper presents a comprehensive review of ML-based and traditional RRA methods in meeting the evolving demands of wireless networks. Literature from relevant articles was selected and analysed to highlight the techniques used, trends, strengths, and limitations. The findings reveal the potential and transformative impact of ML on the future of wireless communications, particularly in achieving the key performance indicators and quality of service expected from 5G and beyond networks. It also shows that research in ML-based RRA methods is at its infancy stage and more research is needed to advance the technology.
A review of PFCP cyber attacks in 5G standalone for robotic telesurgery services
(2024-10) Makondo, Ntshuxeko; Baloyi, Errol; Kobo, Hlabishi I; Mathonsi, TE
The emergence of fifth-generation technology (5G) has revolutionised telecommunication networks, offering enhanced mobile broadband, ultra-reliable (eMBB), ultra-lowlatency communications (uRLLC), and massive machine-type communication (mMTC) service classes. This breakthrough has garnered significant attention and investment worldwide, driving innovation and growth in the digital era. However, the adoption of cloud-based 5G core (5GC) networks, while offering scalability and deployment flexibility, has posed challenges to meeting stringent latency requirements, particularly for uRLLC services specifically for robotic telesurgery. To address this problem, mobile network operators (MNOs) have turned to edge computing (EC), using the control and user plane separation (CUPS) architecture introduced in the thirdgeneration partnership project (3GPP) release 14 specification. This architecture enables the deployment of the user plane function (UPF) closer to users, reducing latency, and improving quality of service (QoS). However, the deployment of the UPF as a standalone node on the edge of the network exposes the packet forwarding control protocol (PFCP) to cybersecurity attacks, which pose risks to telesurgery services and could even lead to loss of life. In the existing literature, only a few techniques focus on minimising these attacks when the UPF is deployed on the edge of the network far from the 5GC. Therefore, this paper reviews PFCP attacks and explores machine learning (ML) techniques to mitigate these security threats. This paper further provides recommendations and future research directions for mitigating these attacks.
Accelerated design of a conformal strongly coupled magnetic resonance wireless power transfer
(2021-11) Molefi, M; Markus, ED; Abu-Mahfouz, Adnan MI
A Conformal Strongly Coupled Magnetic Resonance (CSCMR) wireless power transfer (WPT) system is a small footprint technology suitable for applications such as small low power sensors and implantable medical devices. These applications require specific WPT systems with certain physical dimensions that complement the size of the device. The design of these systems can be complex and require intense computational resources and long simulation times to conceptualise the optimal WPT system. This paper discusses the system architecture for CSCMR-WPT model. A quicker mathematical analysis to estimate the optimal CSCMR-WPT resonator loops and source/load loops is shown. The results confirm that this method can lead to quicker conceptualisation of a WPT model.
Adaptive interference avoidance and mode selection scheme for D2D-enabled small cells in 5G-IIoT networks
(2024-02) Gbadamosi, SA; Hancke, GP; Abu-Mahfouz, Adnan MI
Small cell (SC) and device-to-device (D2D) communications can fulfill high-speed wireless communication in indoor industrial Internet-of-Things (IIoT) services and cell-edge devices. However, controlling interference is crucial for optimizing resource sharing (RS). To address this, we present an adaptive interference avoidance and mode selection (MS) framework that incorporates MS, channel gain factor (CGF), and power-allocation (PA) techniques to reduce reuse interference and increase the data rate of IIoT applications for 5G D2D-enabled SC networks. Our proposed approach employs a two-phase RS algorithm that minimizes the system's computational complexity while maximizing the network sum rate. First, we adaptively determine the D2D user mode for each cell based on the D2D pair channel gain ratios of the cellular and reuse mode. We compute the CGF for each cell with a D2D pair in reuse mode (RM) to select the reuse partner. Then we determine the optimal distributed power for the D2D users and IoT-user equipment using the Lagrangian dual decomposition method to maximize the network sum rate while limiting the interference power. The simulation results indicate that our proposed approach can maximize system throughput and signal-to-interference plus noise ratio, reducing signaling overhead compared to other algorithms.
Addressing accessibility, affordability and sustainability barriers for broadband internet access and penetration in rural areas
(2023-10) Ngwenya, SO; Heymann, R; Swart, TG; Lysko, Albert A
It has been said that the spread of Information and Communications Technology (ICT) and global interconnectedness has a great potential to accelerate human progress, to bridge the “digital divide” and to develop knowledge societies, in countries where it is available. However, it has also been determined that certain barriers, more prevalent in rural communities, pose an incessant impediment to the access and penetration trends of broadband internet networks and infrastructure in rural areas. These barriers have been identified as ‘accessibility’, ‘affordability’, and ‘sustainability’. It is hereby asserted, therefore, that a concerted effort is required to address these barriers, with great deliberation and haste, if digital inclusion is to be achieved in rural areas. This paper proposes a broadband internet system that is aimed at addressing these barriers, through a proposed model, which is intended to be implemented as a Proof of Concept (PoC) in the rural area of Mbazwana, located in the province of KwaZulu-Natal (KZN), South Africa (SA). The model has been presented to the community, and subsequently followed by a survey to determine the feasibility of the implementation thereof. This paper will present the proposed model, as well as the results obtained from the research conducted.
Adsorption of natural dye (Porphyrin and Pheophytin) molecules on TiO2 (101) anatase surface for improved light harvesting efficiency in dye-sensitized solar cells
(2024-07) Ranwaha, TS; Mathomu, LM; Mapanga, Rapela R; Maluta, NE
Dye-sensitized solar cells (DSSCs) have gained popularity in recent years due to their ability to convert sunlight to photovoltaic energy at a low cost. DSSCs use dye molecules adsorbed on TiO2 semiconductors in nanoarchitecture to absorb photons from the sun. In this study, density functional theory was utilized to investigate the geometric, electrical, and optical features of pheophytin and porphyrin dye, as well as its adsorption behaviour on the (010) TiO2 anatase surface. Generalized gradient approximation (GGA) was used to define the exchange-correlation function within the scheme of Perdew-Burke Ernzerhof (PBE), as implemented in the material studio of the BIOVIA CASTEP module. Pheophytin experienced lower photon absorbance in both the visible and near-infrared regions, resulting in lower efficiency. However, when methanol, ethanol, and water solvents were added to the molecule, the blue shift to the visible region and more photon absorption at a higher oscillating strength increased the visible region's efficiency. The results showed that pheophytin and porphyrin dye molecules can improve DSSC performance by shifting absorption to the near-infrared region, improving visible solar spectrum absorption.
Age invariant face recognition methods: A review
(2021-12) Baruni, Kedimotse P; Mokoena, Nthabiseng ME; Veeraragoo, Mahalingam; Holder, Ross P
Face recognition is one of the biometric technologies that is mostly used in surveillance and law enforcement for identification and verification. However, face recognition remains a challenge in verifying and identifying individuals due to significant facial appearance discrepancies caused by age progression. Especially in applications that verify individuals from their passports, driving licenses and finding missing children after decades. The most critical step in Age- Invariant Face Recognition (AIFR) is extracting rich discriminative age-invariant features for each individual in face recognition applications. The variation of facial appearance across aging can be solved using three methods, namely, generative (aging simulation), discriminative (feature-based) and deep neural networks methods. This work reviews and compares the state-of-art AIFR methods to address the work that has been done to minimize the effect of aging in face recognition application during the pre-processing and feature extraction stages to extract rich discriminative age-invariant features from facial images of individuals (subjects) captured at different ages, shortfalls and advantages of these methods. The novelty of this work lies in analyzing the state-of-art work that has been done during the pre-processing and/or feature extraction stages to minimize the difference between the query and enrolled face images captured over age progression.
Analysis of energy-efficient techniques for SDWSN energy usage optimization
(2020-11) Mathebula, I; Isong, B; Gasela, N; Abu-Mahfouz, Adnan MI
Software-Defined Wireless Sensor Networks (SDWSN) has received significant attention in recent years due to its inherent challenges such as network security, trust management, inefficient energy consumption, and so on. In particular, inefficient energy utilization has remained a core critical challenge as sensor nodes are naturally resource constraint and mostly deployed in unattended environments. This challenge has a direct impact on SDWSN performance and reliability. Albeit, several techniques have been proposed and developed to address the challenge in the traditional WSN and few for SDWSN, optimal efficient energy utilization is yet to be achieved. Therefore, this paper survey some of the energy efficiency techniques reported in the literature. The goal is to gain insights into these techniques, strategies employed, their pros and cons which could be utilized to design an efficient-energy mechanism for SDWSN. The findings obtained show that the existing approach ensures efficient energy utilization and improve network performance. While some have routing protocol and security mechanisms, fault tolerance, and battery fault detector mechanisms are not incorporated.
Analysis of hierarchical cluster-based energy-aware routing protocols in WSNs for SDWSN application
(2021-12) Molose, RSS; Isong, B; Dladlu, N; Abu-Mahfouz, Adnan MI
Energy consumption is a typical challenge that is inherent in wireless sensors networks (WSN). Despite the proposals and the development of several routing protocols and Software-defined WSN (SDWSN) introduction to address the challenge and others, optimal energy efficiency is far from being achieved. Therefore, this paper reviewed some of the selected relevant studies based on existing energy-efficient routing protocols to identify the protocols applied and the solutions offered to achieve efficient energy utilization in different applications. Findings show that several different energy-aware routing protocols exist in the WSN with subclasses, each having its strengths and weaknesses. Moreover, empirical evaluation was conducted on 3 commonly deployed hierarchical cluster-based protocols, LEACH, DEEC and TEEN to assess their network lifetime abilities. The results show TEEN protocol outperformed other protocols based on the number of alive nodes with an increase in the number of iterations. We, therefore, recommend the use of these protocols in designing an efficient energy-aware protocol for SDWSN controller placement.
Antenna research directions for 6G: A brief overview through sampling literature
(2021-03) Olwal, TO; Chuku, PN; Lysko, Albert A
Antennas are a critical component in any wireless link. The significance of their role continues to grow, together with the rise of machine-to-machine (M2M) communications and telecommunications starting to move from the fourth generation (4G, or Long-Term Evolution, LTE) to the fifth generation (5G) and beyond 5G (B5G), and also due to the prominence of the convenience promised by the ubiquitous wireless communications. The antennas, or more precisely their large size and appearance, also become a subject of public debate, leading to the need for better antennas, which are can do both: provide high performance and offer visually attractive or nearly invisible/transparent designs. This work reviews several key research trends in antennas to fulfill these demands for the fifth generation of communications (5G) and beyond, especially 6G, and considers novel antenna techniques and designs needed to increase the smartness of the antenna systems and to provide improved beamforming and security.
Antenna research directions for 6G: A brief overview through sampling literature
(2021-03) Olwal, TO; Chuku, PN; Lysko, Albert A
Antennas are a critical component in any wireless link. The significance of their role continues to grow, together with the rise of machine-to-machine (M2M) communications and telecommunications starting to move from the fourth generation (4G, or Long-Term Evolution, LTE) to the fifth generation (5G) and beyond 5G (B5G), and also due to the prominence of the convenience promised by the ubiquitous wireless communications. The antennas, or more precisely their large size and appearance, also become a subject of public debate, leading to the need for better antennas, which are can do both: provide high performance and offer visually attractive or nearly invisible/transparent designs. This work reviews several key research trends in antennas to fulfill these demands for the fifth generation of communications (5G) and beyond, especially 6G, and considers novel antenna techniques and designs needed to increase the smartness of the antenna systems and to provide improved beamforming and security.
Applying phonological feature embeddings for cross-lingual transfer in text-to-speech
(2024-07) Louw, Johannes A; Wang, Z
In this work, we build upon our previous research where we introduced phonological features as input to text-to-speech systems. While the use of phonological features is not a novel concept in our research, our focus in this study is on the comprehensive analysis of the embeddings produced by the encoder model, which we believe offers novel insights into the model’s ability to capture and generalize phonological patterns across languages. Cross-lingual transfer experiments are conducted using both a resource-rich and a resource-constrained language to explore the model’s cross-lingual transfer capabilities across different linguistic families. The analysis of the embedding vectors produced by the encoder model is conducted using cluster maps to visualize the hierarchical clusters obtained using a clustering procedure. This analysis reveals the model’s learning patterns and provides insights into how phonological features contribute to the model’s ability to handle linguistic diversity and data scarcity.
Approximating a Zulu GF concrete syntax with a neural network for natural language understanding
(2021-09) Marais, Laurette
Multilingual Grammatical Framework (GF) domain grammars have been used in a variety of different applications, including question answering, where concrete syntaxes for parsing questions and generating answers are typically required for each supported language. In low-resourced settings, grammar engineering skills, appropriate knowledge of the use of supported languages in a domain, and appropriate domain data are scarce. This presents a challenge for developing domain specific concrete syntaxes for a GF application grammar, on the one hand, while on the other hand, machine learning techniques for performing question answering are hampered by a lack of sufficient data. This paper presents a method for overcoming the two challenges of scarce or costly grammar engineering skills and lack of data for machine learning. A Zulu resource grammar is leveraged to create sufficient data to train a neural network that approximates a Zulu concrete syntax for parsing questions in a proof-of-concept question-answering system.
Assessing the effectiveness of 4IR strategy on South African township economy: Smart Township perspective
(2021-12) Mathibe, Motshedisi; Mochenje, Tonderai; Masonta, Moshe T
As the Coronavirus disease (COVID-19) pandemic changed how people and business interact, its impact severely impacted the global economy. The government imposed heavy lockdown regulations further damaged an already struggling informal sector economy as many livelihoods and informal businesses within the townships came to a halt. Though the South African government introduced various financial relief through the R500 billion support grant, there are township entrepreneurs who could not access the government grant to sustain their everyday business operations. Despite the invisibility of the informal sector, it is considered to contribute about 5 – 30% of the Gross Domestic Product (GDP). Hence, the need to develop a smart township strategy for the township economy will better prepare the township entrepreneurs with a creative and innovative way to survive any future pandemics. This chapter aims to identify factors and enablers that can catalyse critical and innovative thinking to safeguard the township economy and assess the maturity level of the national fourth industrial revolution (4IR) strategy towards building on a South African smart township economy. A digital infrastructure value chain model for building the smart township is proposed. The model also identified digital skills training and upskilling township entrepreneurs to reposition and align themselves within the smart township ecosystems. Furthermore, the chapter proposed key recommendations on how the policy makers could leverage on the 4IR strategy in enabling the township entrepreneurs to actively participate in the digital economy. The objective and significance of this chapter is to assist the policy makers (at provincial and local government level) in understanding both theoretically and practically technological innovation strategies and methods to adopt post-pandemic.
Automatic number plate recognition for remote gate automation: An edge computing approach
(2021-06) Le Roux, E; Ndiaye, M; Abu-Mahfouz, Adnan MI
The article proposes a gate automation system based on automatic number plate recognition (ANPR). The system uses a convolutional neural network (CNN) to identify and classify each number plate which is then compared to a database of authorized numbers before a gate is opened. Once the gate is opened an extra feature of alerting the farm settlement owner via SMS and LoRa is added. What this work tries to do is demonstrate the ability of an IoT-edge device to perform complex image processing computations simply by adding more computing resources at the edge. Edge computing suggests adding a powerful edge device to support edge devices however, we envision the extra computing power can be extended and embedded in the edge devices themselves. To demonstrate this we equip our edge device system with a single board computer to provide the needed computing resources for ANPR.
AwezaMed: A multilingual, multimodal speech-to-speech translation application for maternal health care
(2020-07) Marais, Laurette; Louw, Johannes A; Badenhorst, Jacob AC; Calteaux, Karen V; Wilken, Ilana; Van Niekerk, Nina; Stein, Glenn
The language contexts of multilingual developing countries such as South Africa are often characterised by communication challenges resulting from language differences. AwezaMed is a multilingual, multimodal speech-to-speech translation application for the health care domain, which was designed to assist in bridging communication barriers and mitigate the risks of miscommunication. The application focuses on the domain of maternal health care. It uses English as source language and Afrikaans, isiXhosa and isiZulu as target languages to enable health care providers to communicate with patients in their own language. It incorporates automatic speech recognition, machine translation and text-to-speech to deliver speech-to-speech translation functionality in a scalable way via a REST API to an Android mobile application. It is being piloted at various health care facilities across South Africa.

Browse

Browsing Conference Publications by browse.metadata.cluster "Next Generation Enterprises & Institutions"

Results Per Page

Sort Options