Cameras provide detailed visual information but struggle in turbid water, while sonar can operate in low visibility but often provides less detailed information about objects and their surroundings. Recent advances in computer vision, deep learning and multimodal sensor fusion have improved underwater perception.
However, existing methods often focus on individual sensing technologies or predefined combinations of sensors. A key research challenge is enabling robots to recognise when sensor information becomes unreliable, adapt how they collect and combine observations, and maintain an accurate understanding of their surroundings as conditions change.
This PhD will investigate adaptive multimodal sensing and AI-driven scene understanding to improve the perception capabilities of autonomous underwater robots.
The research will explore how robots can intelligently combine visual, acoustic and other complementary sensor information while adapting their sensing strategies to changing environmental conditions.
The research will focus on: The project will combine machine learning, computer vision, multimodal AI and robotic perception. Research methods will be developed and evaluated using underwater datasets, simulation and controlled experiments, comparing adaptive approaches with conventional sensing and fusion techniques.
The expected outcome is a new framework for adaptive underwater perception that improves scene understanding and sensing reliability under challenging conditions.
The research will contribute towards more capable autonomous underwater robots, with potential applications in subsea inspection, marine infrastructure monitoring and environmental exploration.
The project will be supervised by Dr Ali Rohan, Lecturer in Digital Engineering and Transformation at Loughborough University, whose research focuses on artificial intelligence, computer vision, multimodal perception and autonomous robotic systems. 94% of Loughborough’s research impact is rated world-leading or internationally excellent. REF 2021