Browsing by Author "Mgawe, Bonny"

Now showing 1 - 4 of 4

Augmented Reality-Based Outdoor Navigation System for Urban Firefighting: A Case of Dar es Salaam, Tanzania
(Indian Society for Education and Environment (iSee), 2025-08-11) Sammadile,Zulfa; Luwemba,Godfrey; Mgawe, Bonny; Nyambo,Devotha
Objective: This study aimed to develop an Augmented Reality (AR) based system for outdoor navigation to aid fire truck drivers in reaching incident areas during emergency response in Dar es Salaam, Tanzania. It involved developing three modules: an emergency incident reporting mobile application to improve incident localization, a route planning module to enable the visualization of routes and their conditions, and the AR navigation module to deliver real-time AR-based navigation instructions. Methods: Web AR technologies, specifically WebXR and Three.js, were employed to render AR navigation instructions based on a preselected route from the route planning module, which was developed using the Google Maps JavaScript, Places, and Routes APIs, and complemented with an added layer of house data by leveraging QGIS. Ionic Framework was used to develop the emergency incident reporting platform. The system’s effectiveness was evaluated through User Acceptance Testing (UAT) with 45 Dar es Salaam civilians and 5 fire brigade officers. Findings: The developed system improves incident localization through automatic location sharing via the emergency incident reporting platform and searchable address data via the custom geospatial dataset. It also improves the clarity of navigation instructions through its AR navigation module and routing decisions with its route planning module. During system validation, 91% of the participating civilians completed the emergency incident reporting task without prior training. Furthermore, all the participating officers reported clearer navigation instructions for the AR component, with particular appreciation for the route planning module. Although no direct comparison with existing methods was conducted, validation results revealed the system’s potential in improving emergency response in Dar es Salaam. Novelty: This study presents a context- aware AR navigation system integrated with a custom geospatial dataset that lacks in most previous studies and existing solutions. This integration enhances incident localization, route planning, and navigation, essential factors in improving emergency response.
Dataset: Optimizing LoRaWAN Throughput in Maritime Environments Through Adaptive Coding and Modulation in Rayleigh Fading Channels
(Zenodo, 2025) Lyimo, Martine; Mgawe, Bonny; Leo, Judith; Dida, Mussa; Michael, Kisangiri
This dataset supports the article “Optimizing LoRaWAN Throughput in Maritime Environments through Adaptive Coding and Modulation under Rayleigh Fading”. It includes simulation outputs and MATLAB source code for reproducing all figures and results in the study. Files include throughput, PER, energy efficiency, spectral efficiency, and an ACM algorithm function.
Enhancing detection of common bean diseases using Fast Gradient Sign Method–trained Vision Transformers
(Frontiers in Artificial Intelligence, 2025-08-06) Mwaibale, Upendo; Mduma, Neema; Laizer, Hudson; Mgawe, Bonny
Common bean production in Tanzania is threatened by diseases such as bean rust and bean anthracnose, with early detection critical for effective management. This study presents a Vision Transformer (ViT)-based deep learning model enhanced with adversarial training to improve disease detection robustness under real- world farm conditions. A dataset of 100,000 annotated images augmented with geometric, color, and FGSM-based perturbations, simulating field variability. FGSM was selected for its computational efficiency in low-resource settings. The model, fine-tuned using transfer learning and validated through cross-validation, achieved an accuracy of 99.4%. Results highlight the effectiveness of integrating adversarial robustness to enhance model reliability for mobile-based plant disease detection in resource-constrained environments.
Fine-tuned YOLO-based deep learning model for detecting malaria parasites and leukocytes in thick smear images: A Tanzanian case study
(Elsevier, 2025-09) Lufyagila, Beston; Mgawe, Bonny; Sam, Anael
Malaria remains a serious public health concern in developing countries, where accurate diagnosis is critical for effective treatment. Reliable and timely detection of malaria parasites and leukocytes is essential for precise parasitemia quantification. However, manual identification is labor-intensive, time-consuming, and prone to diagnostic errors—particularly in resource-limited settings. To address this challenge, this paper proposes a fine-tuned deep learning model for detecting malaria parasites and leukocytes in thick smear images. The model is based on the YOLOv10 and YOLOv11 object detection architectures, each independently trained, validated, and evaluated on a custom-annotated dataset collected from hospitals in Tanzania to ensure contextual relevance. A fivefold cross-validation, followed by statistical analysis, was used to identify the best-performing model. Results demonstrate that the optimized YOLOv11m model achieved the highest performance, with a statistically significant improvement (p < .001), attaining a mean mAP@50 of 86.2 % ± 0.3 % and a mean recall of 78.5 % ± 0.2 %. These findings highlight the potential of the proposed model to enhance diagnostic accuracy, support effective parasitemia quantification, and ultimately reduce malaria-related mortality in resource-limited settings.