A Low-Cost Synthetic Aperture Sonar System for Small Agile Vehicles

Abstract

Synthetic Aperture Sonar (SAS) is a coherent acoustic imaging technique capable of producing high resolution, detailed images of underwater scenes. It is of particular use in seafloor mapping operations and detection of unexploded ordinance (UXO), especially for the latter where there is a significant safety benefit to operating at range. The acquisition of detailed imagery is of potential use for target classification algorithms and small object detection, and SAS is becoming more widespread as the technique evolves.

Conventional SAS systems are built on large, custom-built platforms designed to follow linear paths. This means they are typically non-agile and expensive. While this is acceptable for offshore applications, the size prohibits their use in inshore, shallow environments such as rivers and lakes, making SAS systems non-existent in civilian applications. Compounding this problem is the expense associated with building such custom systems, limiting them to use in military trials and high-end research.

This thesis presents a solution by repurposing two pieces of Commercial Off-The-Shelf (COTS) hardware: The Picotech forward looking sonar system and the Blue Robotics remotely operated vehicle (ROV). This results in a small, low-cost, agile platform carrying a functional SAS system, for which several novel techniques were developed. A design methodology based on the Cramer-Rao Lower Bounds (CRLBs) of underwater navigation is presented, and applied to the COTS hardware, identifying an ideal operating region. Subsequently, an algorithm is developed to compensate for the Picotech's curved projection surface, and a novel micro-navigation algorithm, Sector-Scan Beam Correlation (SSBC) is proposed. These are shown to result in improved navigation over existing methods when used in simulations. Experiments performed with the Picotech hardware onboard a small BlueRobotics ROV in local swimming pools demonstrate that the use of SSBC results in clearer images of the target field. Improvement is shown over existing navigation methods, with significant resolution increase over conventional inshore imaging methods.

Date of Award	29 Mar 2023
Original language	English
Awarding Institution	University of Bath
Sponsors	Mr Gareth Hamilton-Fletcher
Supervisor	Alan Hunter (Supervisor), Nigel Johnston (Supervisor) & Pejman Iravani (Supervisor)