Masters Degrees (Electrical and Electronic Engineering)
Permanent URI for this collection
Browse
Recent Submissions
Now showing 1 - 5 of 917
- ItemA cashless payment platform for minibus taxis(Stellenbosch : Stellenbosch University, 2023-03) Tenderere, Kudzai; Booysen, Thinus; Visagie, Lourens; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: The South African Taxi Recapitalisation Program determined that electronic/cashless fare collection systems must be installed in every minibus taxi. However, the past and present initiatives to implement cashless fare collection systems have been focused on using Euro Master Visa (EMV) cards. The problem with these cards is that they exclude part of the population which has been historically marginalised and excluded by financial i nstitutions, m ainly b anks. T his h as h ighlighted t he n eed f or a more inclusive cashless fare collection system in the minibus taxi industry. The cashless fare collection system is also meant to gather data on passenger movement for infrastructure and planning purposes. It was determined that mobile phones have enjoyed a significant p enetration rate into the African and South African population and, as such, mobile money is the best financial technology, particularly for the historically marginalised and financially excluded populations. It was assumed that minibus taxi passengers have Bluetooth devices switched on when they travel. It was determined that social media applications, particularly WhatsApp and Telegram, have become the most used platforms to communicate. Therefore, a decision was made to implement a cashless fare collection system in the form of a Telegram Chatbot: the system would add passenger detection and tracking using Bluetooth. The system for passenger detection and tracking using Bluetooth was designed, implemented, and tested on a couple of local minibus taxi trips. The data gathered from the local trips was analysed to determine the reliability of using Bluetooth for passenger detection and tracking. Generally, Bluetooth use in minibus taxi passengers was neither available nor reliable enough to be used for passenger tracking and identification. The cashless fare collection system was developed without the Bluetooth passenger tracking and identification subsystem.
- ItemMarkerless versus marker-based 3D human pose estimation for strength and conditioning exercise identification.(Stellenbosch : Stellenbosch University, 2023-03) Deyzel, Michael; Theart, Rensu; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: This research aims to contribute to the engineering of a virtual AI-based personal trainer that leverages modern advancements in computer vision. An artificially intelligent personal trainer has the potential to identify, record and assess performances of strength and conditioning exercises where a human trainer is not available. By viewing users through cameras, it could perform 3D human pose estimation to extract the motion of key joints through space. This results in a sequence of skeletons that is a compact representation of the exercise action being performed. In this thesis, we contribute to this concept by developing a motion capture system that records a subject and reconstructs their 3D pose. This can be called markerless motion capture to distinguish it from optical marker-based motion capture used in research and industry. Markerless solutions have the potential to make motion capture more accessible because they are non-intrusive, easy to use and affordable. This would be especially valuable in medical settings where clinicians use motion capture for the diagnosis and treatment of neuromuscular disorders. We used both a research-grade marker-based motion capture system and our markerless system to collect a dataset of motion capture samples of seven different classes of strength and conditioning exercises. We then compared their 3D pose reconstructions to evaluate markerless methods as a means to replace traditional marker-based methods. We studied the errors that must be addressed to make markerless technology truly accessible. From these investigations, we designed strategies that refine markerless pose reconstruction by factoring in the natural expected trajectory of bodily joints. With our most sophisticated method, we could reduce the top 10% error of the markerless motion capture by more than 25%. From the skeleton sequences of different exercises we captured, we developed a skeleton based exercise recognition system using deep learning models. We used a powerful graph convolutional network (GCN) architecture to learn spatial-temporal features for action identification. First, we explored transfer learning by pre-training on a large skeleton-based action dataset, which achieves perfect or near-perfect classification accuracy on our seven exercise classes. Secondly, we explored the more challenging task of one-shot action recognition. This will be a more useful exercise identification system since enrolment of exercises will only require one example of an exercise. We used the GCN model as a feature extractor to learn a metric that projects similar actions closer together in an embedding space and dissimilar actions further apart. Our model achieved a classification accuracy of 87.4% on the seven never-before-seen exercise classes. Our research proves that a markerless motion capture system is sufficient for the capture of 3D pose for application where accuracy and consistency are not of utmost importance, such as for exercise identification, but that more research and development is required before markerless methods can replace the traditional maker-based motion capture used in clinical settings.
- ItemDesign of a series articulated bipedal robot capable of agile and transient maneuvers(Stellenbosch : Stellenbosch University, 2023-03) Weiss, Nathan; Fisher, Callen; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: For legged robots to effectively emulate the dynamic maneuverability, mobility and agility presented by animals in nature, highly dynamic and robust legged locomotive systems are required. In achieving transient legged mobility, legged robots are capable of employing static and dynamically stable motions, including walking, running and jumping, to navigate various topographic terrains and overcome obstacles in unknown environments. However, due to the numerous complexities and non-linearities involved in legged locomotion; researchers in the past have struggled to produce robotic systems that embody the same level of dexterity and maneuverability seen by their biological inspiration. The aim of this thesis was to design and develop a series articulated bipedal robot capable of performing agile and transient maneuvers. In accomplishing this objective, the designed robots would serve as a platform for future research candidates at Stellenbosch University to explore legged locomotive compliance on various terrains. However, the main focus of this research involved the investigation and implementation of key design principles, identified through literature to have contributed to the advancements seen in existing legged robots. To aid the design process of the bipdeal robot, named Q-Bert, an analytical analysis was employed to investigate the jumping performance of a two link articulated leg model for various link lengths and actuators. This resulted in the selection of an appropriate link length, along with a Quasi-Direct Drive electric actuation transmission. Thereafter, an iterative mechanical design process was conducted to produce an initial monopedal prototype; while ensuring adequate structural integrity and minimized system mass and inertia. Furthermore, the planar motion of both, the monopod and biped platforms were constrained within the sagittal plane and supported by a developed vertical planarizing cart system. Q-Bert’s dynamic motions were embodied through the implementation of a virtual model controller inspired by Raibert’s control framework. The performance of these dynamic motions were evaluated and verified through a performance metric known as vertical specific agility. This showed the agility of Q-Bert to surpass some existing dynamic robots; however, was unable to compete with the most agile legged systems. The transient capabilities of Q-Bert were compared to long-time-horizon trajectories generated through a trajectory optimisation simulation and verified Q-Bert’s suitability for transient maneuvers. Q-Bert’s verified suitability was achieved through periodic hopping maneuvers that showed the steady-state hopping frequency and height of the robot to align with the simulated trajectories. Lastly, the maximum recorded jumping height attained by Q-Bert successfully surpassed the analytical jumping height determined during the design analysis and validated the robots design process.
- ItemState of charge and state of health estimation for lithium iron phosphate batteries(Stellenbosch : Stellenbosch University, 2023-03) Snyman, Ricket; Strauss, Johann; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: Accurate state of charge (SOC) and state of health (SOH) is important for safe and optimal operation of lithium batteries. This study investigates multiple existing SOC and SOH estimation methods. These methods are analyzed for their advantages and disad- vantages. There is speci_cally focused on the practicality of these methods to implement them in a typical energy storage system with variable operating conditions. The relationship between the saturation charged energy (energy that is charged dur- ing the saturation charge phase) and the SOH of a lithium iron phosphate (LFP) cell is investigated. A SOH estimation method is proposed based on this relationship that is found. This estimation method only requires partial charging data and does not need data of a full charge or discharge cycle. A hardware system is developed to perform charge and discharge cycles on a single LFP cell while monitoring the vital parameters of the cell. This system is used to perform all testing that is required to develop and evaluate the proposed SOH estimation method. The coulomb counting SOC estimation method is investigated to be improved. Various enhancements are made to the existing method to increase its accuracy and practicality. There is speci_cally focused on considering the charge and discharge e_ciencies of a LFP cell while performing coulomb counting. The OCV (open-circuit voltage) SOC estima- tion method is also combined with the coulomb counting method to form a hybrid SOC estimation method. It was found that the improved hybrid SOC estimation method and the newly found SOH estimation method showed similar accuracy when compared to existing methods. But the main bene_t of these new found methods is their practicality for typical LFP energy storage systems. The improved methods proved to be insensitive to their operating conditions as they do not require a very controlled charge or discharge cycle to obtain accurate estimation values for the SOC and SOH.
- ItemHyperspectral / multispectral imaging technology: application for pomegranate fruit internal quality evaluation and bruise detection(Stellenbosch : Stellenbosch University, 2023-03) Okere, Emmanuel; Perold, Willem; Opara, Linus; Tsige, Alemayehu Ambau; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.ENGLISH ABSTRACT: In recent years, consumer demand for fruit and vegetables are increasing due to a shift towards healthier and more sustainable diets. However, fruits and vegetables are highly perishable that providing the market with high-quality but affordable price is challenging. Also, fruit and vegetable diseases, due to fungal pathogens, are major causes of economic loss in agribusiness. There are multiple sources of contamination during preharvest and harvest–postharvest stages of production and particularly for pomegranate fruit. Pomegranate (Punica granatum L.) is undeniably one of the most ancient deciduous fruits in the world with growing increase in its demand due to its nutritional and health benefits. These quality issues have necessitated rapid and efficient quality and freshness monitoring and analysis tool in the postharvest. In the fruit and vegetable industries, quality inspections are mainly manual and mechanical, laborious, time-consuming, costly, and subjective. Hyperspectral imaging (HSI) has emerged as a powerful non-destructive inspection technique in the agricultural, biosecurity diagnostic and food domain recently. HSI is a non-invasive/ non-destructive technique that integrates spectroscopy as well imaging to form one system. This combined feature makes it a powerful tool for fruit\food quality assessment and defect detection, maturity indexing and physicochemical attributes in horticultural products. Therefore, the main objective of this study is to assess the application of hyperspectral/multispectral imaging for predicting the major quality attributes in fresh pomegranate fruit as well detect the presence of bruise or internal defect using artificial neural networks (ANNs). Section I (Chapter 1, 2 & 3) provides background information, discussing the general aim and objectives (General introduction) of the thesis study. It further provides a comprehensive review on recent applications of hyperspectral imaging technology for preharvest and postharvest analysis for biosecurity diagnostics in the fruit industry (Chapter 2) and narrowed down to applications on pomegranate fruit (Chapter 3). It explores hyperspectral imaging architecture, its equipment, image acquisition and data processing. This information is useful for those in the growers/ processing industries and food safety and quality control stakeholders and provides a review of literature on previous work done on different non-invasive techniques for evaluating different processed horticultural products over the last ten years. In Section II (Chapter 4, 5 and 6), hyperspectral imaging technique was investigated to evaluate maturity quality attributes which includes TSS, TA, pH, and colour components (a*, b*, L*, chrome and hue) of intact pomegranate fruit. The ANN prediction models for quality parameters performed well, with correlation coefficients from 0.421 to 0.951. three neural fitting algorithms were compared for prediction performance, LMG algorithm yielded better results for four of the 9 quality attributes accessed. BR gave the best prediction statistics for TA (R2=0.852, MSE=0.024), and b* (R2=0.951, MSE=3.923). The VNIR spectral was applied to build model using 6 effective wavelengths. This research study has demonstrated that hyperspectral imaging technique in combination with artificial neural network has the potential to predict maturity quality attributes of pomegranate fruit. Further in section II, two spectral ranges of VNIR and SWIR were deployed in the hyperspectral imaging technique to detect the presence of early bruise development of “Wonderful” pomegranate fruit, as well as classify bruit based on different levels of bruise severity. Scanned images were explored, and spectral data extracted for two surface area of interest (ROI and WF). ANN classification model showed model to be able to detect bruise immediately after occurrence to an accuracy of 90%. Both methods of data extraction are good enough to detect the early bruise damage which is invisible to the naked eye. The results confirm hyperspectral imaging technique combined with machine learning methods (ANN) to be an effective technique for early bruise detection. For bruise severity study, both SWIR and VNIR data yielded highly accurate classification results ranging from 80% - 96.7%. The overall average classification accuracy achieved was 93.3% for model to distinguish fruits dropped at 100cm and 90% for fruit dropped at 60cm height for the VNIR camera. Section III (Chapter 7) presents a general discussion on the results and key findings of the different chapters of the thesis. It integrates the results from previous chapters. It highlights the important practical contribution of this thesis towards successful non-destructive evaluation of intact pomegranate fruit.