College of Engineering

A Master thesis was discussed in Department of Computer Engineering/College of Engineering at University of Mosul entitled “Smart Unmanned Robot to Aid Human for Critical Tasks” submitted by the student (Ali Fathel Rasheed) on Thursday, Feb. 20, 2025.

The main objective of this study was to design and implement a prototype ROV capable of underwater monitoring, deep-sea exploration, and assisting in river disaster rescues. This robot aims to perform visual inspections at reduced costs while maintaining structural integrity and functionality. The prototype was designed using SolidWorks CAD and Fusion 360 software (see Appendix A), prioritizing hydrodynamic efficiency and construction simplicity. Materials such as acrylic tubes were selected for their durability and transparency to protect internal electrical components, while PLA was used for 3D-printed structures and floats. The vehicle is equipped with a high-quality camera, LED lighting for dark environments, and a Category 5e multi-pair UTP cable to establish a ground station connection.
Data were generated from controlled experiments where convolutional neural networks (CNNs) were trained on a comprehensive dataset of underwater organisms, achieving 99% accuracy in pure water. To assess the impact of turbidity, kaolin clay was introduced at varying concentrations (21 g/m³, 32 g/m³, …,and 53 g/m³). Results indicated a gradual decline in recognition accuracy, recorded at 98.615% at 21 g/m³, 98.441% at 32 g/m³ , and 98.179% at 53 g/m³, demonstrating the adverse effect of turbidity on object recognition.
Findings suggest that while CNNs function effectively in pure and slightly turbid water, further technological advancements are required to maintain high accuracy in highly turbid environments.