Engineering and Technology

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 20 of 158
  • Item
    Library and Information Services Provision during COVID 19: Ernest Cook Ultrasound Research and Education Institute Library-Mengo Hospital, Uganda’s Perspective
    (4th I-LISS International Conference - IIC 2021, 2021) Kutyamukama, Gitta Alice; Kaddu, Sarah; Abubakar, Mohammed
    COVID-19 pandemic has become a global threat and has established a fear among the mankind. Uganda reported its first case of COVID-19 on the March 21, 2020 (Coronavirus Disease, 2020). A 36-year-old businessman from Kampala, Uganda’s capital, who had travelled to Dubai, United Arab Emirates (UAE) in a healthy condition returned with fever and flu-like symptoms to Entebbe International Airport (EIA) and he was tested positive for SARS-CoV-2 (Coronavirus Disease, 2020). Consequently, individuals who had been to UAE two weeks prior to the first case were traced by Ministry of Health (MoH) Uganda and subjected to institutional quarantine. In the following two weeks (March 21 to April 5), there was a rapid rise in the number of cases to 52, most of whom were imported cases from institutional quarantine ( MOH,Uganda 2020). With the spread of Corona Virus, Uganda was put on lock-down, travel was put to a halt, and Conferences were cancelled, higher institutions of learning, colleges and schools were equally shut-down, the Libraries particularity those in academic institutions faced with unique situations just as their mother institution. It became hard and so difficult for academic libraries to take decisions on how and which library and information services to provide to users amidst full closure of education institutions. In Uganda, normally, library and information services are offered physically and interactively, offer counter interactions with clients and books, Record Centres – offer counter interactions with clients and records. All these are facing hardships. As a way of easing on the lock down and ensuring continuity of learning among students, the government of Uganda recommended online learning to ensure continuity for learners. Though already challenged, such a recommendation piled more pressure on the already financially and structurally strained education system. It left a lot of questions and challenges on how to cope with this new dimension of learning albeit ill preparedness and exploration of new grounds for learners of the education system. The Library is taken to be a competent authority for providing access to authoritative, current and timely information to support learning and research while leveraging innovations of the information age brought about by ICTs.
  • Item
    Gender Bias Evaluation in Luganda-English Machine Translation
    (Association for Machine Translation in the Americas, 2022-09-07) Wairagala, Eric Peter; Mukiibi, Jonathan; Babirye, Claire; Nakatumba-Nabende, Joyce; Katumba, Andrew; Ssenkungu, Ivan
    We have seen significant growth in the area of building Natural Language Processing (NLP) tools for African languages. However, the evaluation of gender bias in the machine translation systems for African languages is not yet thoroughly investigated. This is due to the unavailability of explicit text data available for addressing the issue of gender bias in machine translation. In this paper, we use transfer learning techniques based on a pre-trained Marian MT model for building machine translation models for English-Luganda and Luganda-English. Our work attempts to evaluate and quantify the gender bias within a Luganda-English machine translation system using Word Embeddings Fairness Evaluation Framework (WEFE). Luganda is one of the languages with gender-neutral pronouns in the world, therefore we use a small set of trusted gendered examples as the test set to evaluate gender bias by biasing word embeddings. This approach allows us to focus on Luganda-Engish translations with gender-specific pronouns, and the results of the gender bias evaluation are confirmed by human evaluation. To compare and contrast the results of the word embeddings evaluation metric, we used a modified version of the existing Translation Gender Bias Index (TGBI) based on the grammatical consideration for Luganda.
  • Item
    Green Computing Knowledge among Students in a Ugandan University
    (IEEE., 2016-11-22) Semakula, Isa; Samsuri, Suhaila
    Humans, including students, are partly responsible for rising levels of global warming through unsustainable computing practices. It is imperative that students participate in the global drive to reduce global warming through activities like green computing which advocates for sustainable use of computing resources to reduce energy consumption and carbon emissions. The study was conducted to establish green computing awareness levels of students from Islamic University in Uganda (IUIU) and also determine if such levels were influenced by computer experience. Results from a purposive sample of 452 students indicated that students' levels of green computing awareness were low, and computer experience influenced two of the three dimensions of perceived knowledge of green computing.
  • Item
    A Literature Review on Sampling Techniques in Semiconductor Manufacturing
    (IEEE Transactions on Semiconductor Manufacturing, 2013) Nduhura-Munga, Justin; Rodriguez-Verjan, Gloria; Dauz ´ ere-P ` er´ es, Stephane; Yugma, Claude; Vialletelle, Philippe; Pinaton, Jacques
    This paper reviews sampling techniques for inspection in semiconductor manufacturing. We discuss the strengths and weaknesses of techniques developed in the last last 20 years for excursion monitoring (when a process or machine falls out of specifications) and control. Sampling techniques are classified into three main groups: static, adaptive, and dynamic. For each group, a classification is performed per year, approach, and industrial deployment. A comparison between the groups indicates a complementarity strongly linked to the semiconductor environment. Benefits and drawbacks of each group are discussed, showing significant improvements from static to dynamic through adaptive sampling techniques. Dynamic sampling seems to be more appropriate for modern semiconductor plants.
  • Item
    The relationship between the Old Case Loads and New Case Loads in Nakivale refugee settlement in Southwestern Uganda
    (Journal of Eastern African Studies, 2016)
    Forced migration from Rwanda has become a common trend. It has been there since time immemorial and is still continuing at varying magnitudes. Scholarly literature indicate that over 100,000 Rwandan refugees crossed into Uganda and other neighbouring countries of Burundi, Tanzania and Zaire (the present day Democratic Republic of Congo) in early 1960s seeking asylum following violent political and social changes in Rwanda that erupted in 1959. Rwanda was a former colony of Germany before the First World War which later came under the Belgian administration. It is comprised of three ethnic groups, Tutsi, Hutu and Twa, with the Hutu taking 85%, followed by the Tutsi at 14% and the Twa at 1% of the total population. Interestingly, they all speak the same language; Ikinyarwanda. Prior to colonization, the Tutsi, although smaller in number, were the ruling ethnic group over the majority Hutu. In 1959, as Rwanda was warming up to attain independence, tensions started erupting between the Tutsi and Hutu over who should rule the new country. It is believed that although the Tutsi were few in numbers, they were more elite and intelligent compared to the majority Hutu. The Hutu, having been mistreated and under looked during the reign of Tutsi, they were determined to supersede them this time round. As the tensions heightened, many Tutsi were killed and the remnants fled to neighbouring countries. To some scholars, this was blamed on the poor management by the Belgians who are believed to have dragged their feet at the beginning and later moved too fast without proper preparation for decolonization of the local communities. The Belgian colonial masters put in place identity cards that classified individual Rwandans as either Hutu or Tutsi which meant permanent racialised divisions in favour of the ruling Tutsi by then1.
  • Item
    ADSMS: Anomaly Detection Scheme for Mitigating Sink Hole Attack in Wireless Sensor Network
    (IEEE, 2017) Yasin, N. Mohammaed; Balaji, N.; Sambasivam, G.; Basha, M. S. Saleem; Sujatha, P.
    In past years, mobile ad hoc networks (MANETs) widespread use in many applications, including for some mission-critical applications, and has become one of the major concerns such as security MANETs. MANETs due to some unique characteristics, prevention methods are not alone enough to protect them need; Therefore, the detection is possible for an attacker to breach system security, such as the need to add another before. In general, traditional wireless networks, intrusion detection techniques are not well suited for MANETs. In this case, to protect it from attacks MANETs is important to develop more efficient methods of intrusion detection. With improvements in technology and cutting hardware costs, we MANETs expanding into industrial applications are also witnessing a current trend. To cope with such a trend and we believe strongly that it was important for its potential security issues. In this paper, we propose new intrusion detection and specially designed MANETs Improved Acceptance acknowledgment (EAACK) to activate the digital signature system. Compared to contemporary approaches, while not greatly affect network shows EAACK certain conditions demonstrates the high detection rates of malicious behaviour
  • Item
    A Normalized Approach for Service Discovery
    (Procedia Computer Science, 2015) Sambasivam, G.; Ravisankar, V.; Vengattaraman, T.; Baskaran, R.; Dhavachelvan, P.
    In today’s world web services are the known perception to all the users who uses the internet. The Web Service process involves service discovery, selection and ranking. Discovery is the process of matchmaking of user query with advertisements in the repository. Our motivation is to develop a model for web service discovery with the combined approach of service selection and ranking. In this paper, we have proposed a technique for web service discovery process, combining the keyword search and semantic search and ranking the services. The implementation results show that the proposed model performs better for the web service discovery process.
  • Item
    Mira: an Application Containerization Pipeline for Small Software Development Teams in Low Resource Settings
    (Development Engineering, 2022) Mwotil, Alex; Bainomugisha, Engineer; Araka, Stephen G.M.
    Cloud native applications leverage Development and Operation (DevOps), microservice architectures and containerisation for primarily availability, resilience and scalability reasons. Small developer teams in low resource settings have unique DevOps needs and harnessing its principles and practices is technically challenging and distinctly difficult in these contexts. We conducted a survey with professional developers, students and researchers situated and working in a low resource setting and the results indicate that these principles and practices are relatively new. In application containerisation, an operating system virtualisation method that can significantly optimize the use of computing resources, the respondents indicated challenges in the process steps. Particularly, small developer teams in low resource settings require custom tools and abstractions for software development and delivery automation. Informed by the developer needs, we designed and developed a custom automated containerisation pipeline, mira, for a managed cloud native platform situated in a low-resource setting. We validate mira against 6 major application frameworks, tools and/or languages and successful deployment of the resultant applications onto a cloud native platform.
  • Item
    Hydraulic conductivity predictive model of RHA-ameliorated laterite for solving landfill liner leachate, soil and water contamination and carbon emission problems
    (International Journal of Low-Carbon Technologies, 2022) Onyelowe, Kennedy C.; Ebid, Ahmed M; Baldovino, Jair de Jesús Arrieta; Onyia, Michael E.
    The environment is seriously being affected by the leachate release at the unconstructed and badly constructed waste containment or landfill facilities around the globe. The worst hit is the developing world where there is little or totally no waste management system and facilities to receive waste released into the atmosphere. This research work is focused on the leachate drain into the soil and the underground water from landfills, which toxicifies both the soil and the water. Also, the construction of the liner or barrier with cement poses serious threat to the environment due to oxides of carbon release and this research also took this into account by replacing the utilization of cement with rice husk ash (RHA), which has proven to have the potentials of replacing cement as a supplementary binder. Laboratory tests were conducted to determine the hydraulic conductivity (K) of lateritic soil (LS) ameliorated with different dosages of RHA. Other hydromechanical properties of the treated blend were studied and multiple data were generated for the artificial neural network (ANN) back-propagation (-BP), genetic algorithm (GA) and gradual reducing gradient (GRG), genetic programming (GP) and evolutionary polynomial regression (EPR) prediction exercises. Results show that the LS was a poorly graded A-2 sandy silt soil, which was subjected to three different compaction energies with the minimum of the British standard light (BSL) and derived k of 6.95E-10, 50.75E-10 and 32.33E-10 for BSL, west African standard and British standard heavy, respectively. The RHA addition improved the studied properties of the ameliorated LS. Out of the five models, the ANN-GRG outclassed others with a performance of 99% with minimal error compared with the rest. Potentially, this research has shown that RHA with a pozzolanic chemical moduli of 81.47% can replace cement in the construction of ecofriendly and more efficient landfills and waste containemnt barriers to save the soil and the underground water as well as the environment from leachate contamination and carbon emissions.
  • Item
    Energy Efficiency and Spectral Efficiency Trade-off of a Novel Interference Avoidance Approach for LTE-Femtocell Networks
    (IEEE., 2012) Wang, Siyi; Turyagyenda, Charles; O’Farrell, Tim
    The trade-off between energy efficiency and downlink spectral efficiency for LTE-femtocells can be significantly balanced by a novel interference management technique in the presence of the interference from the co-channel outdoor micro-cell and the neighbouring femtocell access points. The simulation results have been demonstrated to be meaningful in the context of considering capacity saturation of realistic modulation and coding schemes rather than theoretical Shannon's equation. The paper shows that the radio-head and operational improvement of up to 12% and 3% can be achieved. Moreover, it has been shown that the improvement does not significantly degrade the user's data rate and the proposed scheme can be implemented in unplanned self- organising networks.
  • Item
    Security Analysis of Mobile Edge Computing in Virtualized Small Cell Networks
    (Springer International Publishing., 2016) Vassilakis, Vassilios; Chochliouros, Ioannis P.; Spiliopoulou, Anastasia S.; Turyagyenda, Charles; Dardamanis, Athanassios
    Based upon the context of Mobile Edge Computing (MEC) actual research and within the innovative scope of the SESAME EU-funded research project, we propose and assess a framework for security analysis applied in virtualised Small Cell Networks, with the aim of further extending MEC in the broader 5G environment. More specifically, by applying the fundamental concepts of the SESAME original architecture that aims at providing enhanced multi-tenant MEC services through Small Cells coordination and virtualization, we focus on a realistic 5G-oriented scenario enabling the provision of large multi-tenant enterprise services by using MEC. Then we evaluate several security issues by using a formal methodology, known as the Secure Tropos.
  • Item
    Energy Efficiency Comparison of Common Packet Schedulers
    (IEEE., 2014) Turyagyenda, C.; Albeiruti, N.; Zak, M.; Al-Begain, K.
    This paper presents a characterisation of the energy performance of common and well-known packet schedulers in a multi-cell multi-user UTRAN long-term evolution (LTE) system. First, it was shown that the maximum SINR metric is more energy efficient than the round robin metric for both time and frequency domain packet scheduling. Second, it was shown that frequency domain packet scheduling had a greater impact on the energy efficiency of the system than time domain packet scheduling; i.e. RF Energy Reduction Gains (RF-ERG) ranging from 16% to 49% and 2% to 18% respectively.
  • Item
    Long Term Evolution Downlink Packet Scheduling using A Novel Proportional-Fair-Energy Policy
    (IEEE., 2012) Turyagyenda, C.; O'Farrell, T.; Guo, W.
    Inter-cell interference (ICI) is a key limiting factor to the general performance of a multi-cell multi-user radio access network. The channel quality of cell edge users is greatly impaired by ICI owing to the fact that cell edge users are furthest away from the their serving base station and closest to the interfering base stations. As a result the Quality of Service (QoS) and energy efficiency of the E-UTRAN is primarily dependant on the cell edge users. Firstly, we propose a new Time Domain Packet Scheduling criterion that endeavours to reduce the variation in the energy performance, of the users, in a temporal sense. The proposed criteria aims to strike a balance between two user prioritisation criteria that result in energy performance at the two extremes of the energy consumption range. The paper shows that this improves the mean energy efficiency of the E-UTRAN. Secondly, we introduce an energy optimisation algorithm to complement the Time Domain Packet Scheduler. The new energy aware packet scheduling criteria is compared against the established throughput based proportional fair scheduler with uniform power allocation and is shown to produce 20% Energy Reduction Gains (ERG) without compromising the spectral efficiency and QoS performance.
  • Item
    SFBC MIMO Energy Efficiency Improvements of Common Packet Schedulers for the Long Term Evolution Downlink
    (IEEE., 2011) Turyagyenda, C.; O’Farrell, T.; He, J.; Loskot, P.
    It is desirable that energy performance improvement is not realized at the expense of other network performance parameters. This paper investigates the trade off between energy efficiency, spectral efficiency and user QoS performance for a multi-cell multi-user radio access network. Specifically, the energy consumption ratio (ECR) and the spectral efficiency of several common frequency domain packet schedulers in a cellular E- UTRAN downlink are compared for both the SISO transmission mode and the 2×2 Alamouti Space Frequency Block Code (SFBC) MIMO transmission mode. It is well known that the 2×2 SFBC MIMO transmission mode is more spectrally efficient compared to the SISO transmission mode, however, the relationship between energy efficiency and spectral efficiency is undecided. It is shown that, for the E-UTRAN downlink with fixed transmission power, spectral efficiency improvement results into energy efficiency improvement. The effect of SFBC MIMO versus SISO on the user QoS performance is also studied.
  • Item
    Performance Security Trade-off of Network Intrusion Detection and Prevention Systems
    (UK Performance Engineering Workshop and Cyber Security Workshop, 2016) Munir, Rashid; Ahmed, Botan; Al-Mohannadi, Hamad; Mufti, M. Rafiq; Namanya, Anitta Patience; Awan, Irfan
    Security cyber threats are increasing with most companies being overwhelm by the complexity attached to prevention against attacks. Network Intrusion detection and prevention systems (NIDPS) are now a stable in any enterprise network with the purpose of filtering through the network traffic and sniffing for malicious traffic. Given the amount of traffic generated on enterprise networks nowadays, any NIDPS is sure to go through a big number of packets that a need arises for a performance- security trade-off. On any given day, based on the rules used in the NIDPS, the number of alerts it generates are in thousands. This can be quite overwhelming to security analysts who analyse them to understand the cyber threat landscape. Although it is true the more alerts, the higher the probability of detecting malicious traffic, it is also true that alerts require the traffic to go through many rules which can be quite a performance hindrance. This is the paradox plagued by the cyber security community currently. In this paper, we examine 2 scenarios to evaluate the performance security trade-off for the purpose of propose ways of improving the performance while minimising the impact on the security purpose for the NIDPS.
  • Item
    Performance Modelling and Analysis of the Delay Aware Routing Metric in Cognitive Radio Ad Hoc Networks
    (IEEE, 2013) Namanya, Anitta Patience; Pagna-Disso, Jules
    Cognitive Radio Networks have been proposed to solve the problem of overcrowded unlicensed spectrum by using the cognitive ability built in software radios to utilise the underutilised licensed channel when the licensed users are not using it. Successful results from the research community have led to its application to wireless technologies like Ad Hoc networks due to their extensive advantages. Cognitive Radio Ad Hoc networks are a novel technology that will provide a solution to many communication challenges. This paper investigates the end-to-end performance modelling of a link using quality of service parameters; delay vs. link capacity while considering the factors of spectrum management and node mobility of two nodes in tandem representing a hop in Cognitive Radio Ad Hoc networks. We modelled spectrum management and node mobility using the pre-emptive resume priority M/G/1 queuing model and the gated node model respectively. We considered delay aware routing schemes; shortest queue and random probability routing and compared them with the analytical link-capacity for analysis. The study shows that already existing mathematical models can be used as close approximations to analyse the queuing models proposed for Cognitive Radio Ad Hoc Networks.
  • Item
    Detection of Malicious Portable Executables using Evidence Combinational Theory with Fuzzy Hashing
    (IEEE, 2016) Namanya, Anitta Patience; Khan Ali Mirza, Qublai; Al-Mohannadi, Hamad; Awan, Irfan U.; Ferdinand Pagna Disso, Jules
    Fuzzy hashing is a known technique that has been adopted to speed up malware analysis processes. However, Hashing has not been fully implemented for malware detection because it can easily be evaded by applying a simple obfuscation technique such as packing. This challenge has limited the usage of hashing to triaging of the samples based on the percentage of similarity between the known and unknown. In this paper, we explore the different ways fuzzy hashing can be used to detect similarities in a file by investigating particular hashes of interest. Each hashing method produces independent but related interesting results which are presented herein. We further investigate combination techniques that can be used to improve the detection rates in hashing methods. Two such evidence combination theory based methods are applied in this work in order propose a novel way of combining the results achieved from different hashing algorithms. This study focuses on file and section Ssdeep hashing, PeHash and Imphash techniques to calculate the similarity of the Portable Executable files. Our results show that the detection rates are improved when evidence combination techniques are used.
  • Item
    An Edge and Fog Computing Platform for Effective Deployment of 360 Video Applications
    (IEEE., 2019) Rigazzi, Giovanni; Kainulainen, Jani-Pekka; Turyagyenda, Charles; Mourad, Alain; Ahn, Jaehyun
    Immersive video applications based on 360 video streaming require high-bandwidth, high-reliability and low-latency 5G connectivity but also flexible, low-latency and cost-effective computing deployment. This paper proposes a novel solution for decomposing and distributing the end-to-end 360 video streaming service across three computing tiers, namely cloud, edge and constrained fog, in order of proximity to the end user client. The streaming service is aided with an adaptive viewport technique. The proposed solution is based on the H2020 5G-CORAL system architecture using micro-services-based design and a unified orchestration and control across all three tiers based on Fog05. Performance evaluation of the proposed solution shows noticeable reduction in bandwidth consumption, energy consumption, and deployment costs, as compared to a solution where the streaming service is all delivered out of one computing location such as the Cloud.
  • Item
    Prediction of cutting force for self-propelled rotary tool using artificial neural networks
    (Journal of Materials Processing Technology, 2006) Hao, Wangshen; Zhu, Xunsheng; Li, Xifeng; Turyagyenda, Gelvis
    In this paper, a cutting force model for self-propelled rotary tool (SPRT) cutting force prediction using artificial neural networks (ANN) has been introduced. The basis of this approach is to train and test the ANN model with cutting force samples of SPRT, from which their neurons relations are gradually extracted out. Then, ANN cutting force model is achieved by obtaining all weights for each layer. The inputs to the model consist of cutting velocity V, feed rate f, depth of cut ap and tool inclination angle λ, while the outputs are composed of thrust force Fx, radial force Fy and main cutting force Fz. It significantly reduces the complexity of modeling for SPRT cutting force, and employs non-structure operator parameters more conveniently. Considering the disadvantages of back propagation (BP) such as the convergence to local minima in the error space, developments have been achieved by applying hybrid of genetic algorithm (GA) and BP algorithm hence improve the performance of the ANN model. Validity and efficiency of the model were verified through a variety of SPRT cutting samples from our experiment tested in the cutting force model. The performance of the hybrid of GA–BP cutting force model is fairly satisfactory.
  • Item
    Integrated Cross-Layer Energy Savings in a Smart and Flexible Cellular Network
    (IEEE., 2012) Guo, Weisi; Wang, Siyi; Turyagyenda, Charles; O’Farrell, Tim
    A key challenge for mobile operators is how to reduce the operational energy and cost expenditure, whilst meeting the growing demand for throughput. In recent years, individual research techniques have shown that significant savings can be made. The majority of savings are achieved in the signal transmission stage and are obtained under certain modeling conditions and assumptions. How the gains can be combined together to yield higher total operational savings is largely unexplored, especially under a realistic multi-cell multi-user environment. This paper employs an integrated analysis of the cross-layer techniques that reduce energy consumption or improve the spectral- and energy- efficiency tradeoff. The research is part of the key integration process of the MVCE Green Radio (GR) programme, which combines architecture, transmission technique, resource management, and hardware research. The integrated operational energy savings have been shown to be above 90% and the associated cost savings are up to 34%. Furthermore, the paper discusses the impact of machine-learning and energy harvesting on the energy and cost consumption, to create a smart and flexible cellular network.