Back to search

University of Southampton

A Predictive Engine for environmentally-adaptive marine robotic swarm operations

Abstract

dc:description.abstract

Marine robotic swarms offer an attractive means of collecting in situ environmental observations in hazardous and data-sparse regions, including tropical cyclones. In such missions, the marine environment influences both platform performance (e.g., achievable speed and motion response) and the availability of inter-agent communication. These coupled effects complicate mission design and, without suitable models, limit decision-making to conservative planning and largely reactive behaviour. This thesis addresses this challenge by developing lightweight, computationally tractable models intended to support high-throughput simulation studies and, where feasible, onboard decision support.<br/><br/>The thesis introduces a high-performance Predictive Engine whose core component is ASVLite, an open-source simulator designed for scalable ensemble evaluation. ASVLite combines spectral wave synthesis with efficient wave-vehicle interaction modelling, enabling faster-than-real-time execution for multi-vehicle scenarios and providing a practical basis for large parameter sweeps. Computational performance is characterised on both desktop hardware and Raspberry Pi single-board computer to support the swarm-scale motivation. The physical modelling is validated within a limited scope through the development of a wave glider dynamics model and comparison against real-world trans-Pacific voyage data; this validation provides evidence of sensible trend-level behaviour but does not constitute comprehensive seakeeping validation across all degrees of freedom and sea states.<br/><br/>The Predictive Engine is then applied across two operational timescales. For long-horizon strategic planning, it is used to perform a comparative, exploratory evaluation of hand-designed deployment strategies for tropical cyclone monitoring, enabling operators to examine trade-offs between interception likelihood and operational risk under specified environmental conditions. For shorter-horizon communication analysis, the thesis develops a quantitative model demonstrating that wave-induced diffraction can be a dominant contributor to link intermittency for low-profile surface platforms. The analysis explicitly distinguishes between low-rate communication required for health/status and coordination and higher-throughput local exchange required for sharing richer mission products, motivating hybrid satellite--line-of-sight assumptions. A conceptual case study illustrates how predicted link degradation could inform formation-level adaptation, while acknowledging that robust closed-loop swarm control is not experimentally established here.<br/><br/>Overall, the thesis contributes a coherent modelling and simulation framework for environment-aware assessment of marine swarms, integrating lightweight wave generation, platform response modelling (within stated validation limits), and communication reliability analysis. By enabling repeatable, scalable studies of mission sensitivity to environmental conditions, the work supports more informed planning and provides a foundation for future development of proactive, comms-aware swarm autonomy for demanding applications such as tropical cyclone observation.

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
  • Thomas, Toby
Advisors dc:contributor.advisor
  • Tarapore, Danesh
  • Zauner, Klaus-Peter
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

Thomas, Toby. A Predictive Engine for environmentally-adaptive marine robotic swarm operations. doctoral thesis, University of Southampton, 2026. https://doi.org/10.5258/SOTON/PG/T226