Ioannis Polymenis
Autonomous underwater intervention.
Email: i.polymenis@ncl.ac.uk
Project supervisors
Project description
I am assessing the feasibility of physical intervention by an Autonomous Underwater Vehicle (AUV).
Methodology and objectives
My research focuses on infrastructure inspection. I will explore methods of detecting faults that could occur in a variety of underwater structures. I will then suggest necessary actions. I am using Neural Networks (NN) techniques for Computer Vision (CV) in an underwater vehicle. The underwater robot will analyse the acquired information. From this, it should be able to make critical decisions on how to intervene in any scenario to solve and mitigate the problem.
Initially, I will train the system’s image recognition using an in-house developed dataset of images. Next, I will introduce the system to the actual underwater environment with further complexity, such as low visibility. I will then further develop the system to make decisions on how to react or intervenewhen it recognises a fault.
I will carry out virtual simulations to assess the system’s performance and make the necessary adjustments at each stage.
Finally, I will implement the fully developed system on a physical prototype. I will test it in the hydrodynamics laboratory to assess its performance.
Outcomes
The project will develop an algorithm capable of identifying and inspecting underwater objects in a real underwater environment. It will provide appropriate physical interventions when it detects a fault.
Qualifications
- BEng Hons Marine Technology with Marine Engineering, Newcastle University
- MSc Naval Architecture, Newcastle University