Back to search

University of Southampton

Prior-informed perception and adaptive survey planning for autonomous underwater vehicles

Abstract

dc:description.abstract

Autonomous Underwater Vehicles (AUVs) are central to seafloor mapping, habitat monitoring, and environmental assessment, yet limited sensing footprints, imperfect priors, and tight operational budgets make exhaustive coverage infeasible. Since most surveys are still conducted by a single vehicle over regions for which coarse maps already exist, this thesis asks how the prior information already available---remotely sensed maps, the spatial arrangement of observations, and expert survey intent---can be leveraged to interpret and plan sparse surveys.<br/><br/>Location metadata is introduced as a spatial regularisation signal for learning representations from multimodal environmental priors, improving feature quality and habitat classification; this benefit is shown to generalise across multiple self-supervised frameworks and architectures, revealing trade-offs between model capacity and generalisation. To support long-duration deployments, an online clustering framework interprets streaming seafloor imagery in real time with bounded computational cost.<br/><br/>Building on these perception capabilities, an adaptive prior-informed planning framework treats sparse surveying as a controllable feature-space sampling process, combining optimised survey templates that encode expert-defined habitat priorities with information-driven adaptive transits coordinated through a hybrid behavioural architecture. This steers survey effort towards prioritised habitats under a shared mission budget while improving the representativeness of the collected observations, enabling AUVs to conduct efficient, robust, and scientifically meaningful surveys in real-world environments.

Degree

thesis:*
Name dc:type.qualificationname
Ph.D.
Level dc:type.qualificationlevel
doctoral
Grantor dc:publisher.institution
University of Southampton
Year dc:date.issued
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Liang, Cailei
Advisors dc:contributor.advisor
  • Thornton, Blair
  • Turnock, Stephen
ORCID person:identifier.orcid

Rights

Language dc:language
English, eng

Identifiers

dc:identifier.*

Chain of custody

source
Harvested from
University of Southampton
Base URL
eprints.soton.ac.uk/cgi/oai2
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
related terms
citation

Liang, Cailei. Prior-informed perception and adaptive survey planning for autonomous underwater vehicles. doctoral thesis, University of Southampton, 2026. https://doi.org/10.5258/SOTON/01/T018