Matching Items (252)
Description
As part of a United States-Australian Solar Energy Collaboration on a Micro Urban Solar Integrated Concentrator project, the purpose of the research was to design and build a bench-top apparatus of a solar power concentrator thermal storage unit. This prototype would serve to be a test apparatus for testing multiple thermal storage mediums and heat transfer fluids for verification and optimization of the larger system. The initial temperature range for the system to test a wide variety of thermal storage mediums was 100°C to 400°C. As for the thermal storage volume it was decided that the team would need to test volumes of about 100 mL. These design parameters later changed to a smaller range for the initial prototype apparatus. This temperature range was decided to be 210°C to 240°C using tin as a phase change material (PCM). It was also decided a low temperature (<100°C) test using paraffin as the PCM would be beneficial for troubleshooting purposes.
ContributorsLee, William John (Author) / Phelan, Patrick (Thesis director) / Wang, Robert (Committee member) / Barrett, The Honors College (Contributor) / Mechanical and Aerospace Engineering Program (Contributor) / School of International Letters and Cultures (Contributor)
Created2015-05
Description
The thesis explores the avenues of machine learning principles in object detection using TensorFlow 2 Object Detection API Libraries for implementation. Integrating object detection capabilities into ESP-32 cameras can enhance functionality in the capstone dragster application and potential applications, such as autonomous robots. The research implements the TensorFlow 2 Object Detection API, a widely used framework for training and deploying object detection models. By leveraging the pre-trained models available in the API, the system can detect a wide range of objects with high accuracy and speed. Fine-tuning these models using a custom dataset allows us to enhance their performance in detecting specific objects of interest. Experiments to identify strengths and weaknesses of each model's implementation before and after training using similar images were evaluated The thesis also explores the potential limitations and challenges of deploying object detection on real-time ESP-32 cameras, such as limited computational resources, costs, and power constraints. The results obtained from the experiments demonstrate the feasibility and effectiveness of implementing object detection on ESP-32 cameras using the TensorFlow2 Object Detection API. The system achieves satisfactory accuracy and real-time processing capabilities, making it suitable for various practical applications. Overall, this thesis provides a foundation for further advancements and optimizations in the integration of object detection capabilities into small, low-power devices such as ESP-32 cameras and a crossroad to explore its applicability for other image-capturing and processing devices in industrial, automotive, and defense sectors of industry.
ContributorsMani, Vinesh (Author) / Tsakalis, Konstantinos (Thesis director) / Jayasuriya, Suren (Committee member) / Barrett, The Honors College (Contributor) / Electrical Engineering Program (Contributor)
Created2024-05