math.OCAug 27, 2026

Marine Autonomous Vehicle Fleet Scheduling to Maximise Scientific Impact

Authors: Mehdi El Krari, Jonathan Smith, Maria Fox

Organizations: British Antarctic Survey, AI Lab Cambridge, UK

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

The marine science community increasingly relies on Marine Autonomous Vehicles (MAVs) to collect the critical environmental data required to understand global ocean systems. However, as these operations scale, manually routing and planning large autonomous fleets becomes exponentially complex and time-consuming. To address this, we propose a mixed-integer linear programming (MILP) model designed to automate and optimise MAV deployment schedules. The model accounts for strict operational constraints, including battery capacities and time windows for data collection, while aiming to maximise total data collection and minimise both the number of deployed vehicles and their energy consumption. A key novelty of this framework is integrating conventional ship itineraries, allowing MAVs to support vessels with mid-mission battery swapping or accelerated transit between waypoints. Computational experiments demonstrate that the model is highly scalable, solving routing problems for fleets of dozens of MAVs in seconds, and scaling to hundreds of vehicles in only a few minutes. Beyond operational scheduling, the framework serves as a robust simulation tool for evaluating 'what-if' scenarios and analysing the impact of varying parameters on deployment strategies. Finally, the solution generates a suite of visualisations designed to enhance explainability and support strategic decision-making for stakeholders.

Figures & tables

Figure 1: Shoreside deployment from Rothera, Antarctica. (a) Represents example routes from the deployment/recovery location Rothera (R) to each of the research stations (S1, S2 and S3) for Slocum marine gliders. The transit time in days for each route is labelled. (b) Represents the same as panel (a) but for the AL1500.
AUV TypesBase LocationPickup LocationInitial Battery Level (%)
Autonomous PlatformSensors
SlocumCTD+microstructure+ADCPRotheraRothera80
SlocumCTD+microstructure+ADCPRotheraRothera60
SlocumCTD+microstructure+ADCPRotheraRothera100
SlocumCTD+microstructure+ADCPRotheraRothera88
SlocumCTD+ADCPRotheraRothera80
Table 1: Fleet details for deployment around Rothera
Research StationsAUV TypesNumber RequiredTask Duration (days)Earliest StartLatest End
S1Slocum - CTD, Microstructure, ADCP282025-01-132025-02-05
S1Slocum - CTD, ADCP2102025-01-112025-02-10
S1Slocum - CTD, Optode, PAR, Ecopuck2122025-01-162025-03-17
S2Slocum - CTD, Microstructure, ADCP372025-01-212025-03-22
S2Slocum - CTD, ADCP192025-01-312025-03-22
S2Slocum - CTD, Optode, PAR, Ecopuck152025-01-262025-03-17
Table 2: Requested science surveys around Rothera
Figure 2: Timeline of MAV activities for the deployment around Rothera
Figure 3: Hybrid ship and shoreside deployment around South Georgia and the Falklands. The red dashed line with arrows represents the pre-defined route of the vessel returning from the Weddell Sea before heading north, travelling through waypoints shown by circles. The research stations are shown by the rectangles.
AUV TypesBase LocationPickup LocationInitial Battery Level (%)
Autonomous PlatformSensors
SlocumCTD+microstructure+ADCPFalklandsRDV51
SlocumCTD+microstructure+ADCPFalklandsCheckpoint 143
SlocumCTD+microstructure+ADCPSouth GeorgiaCheckpoint 1100
SlocumCTD+microstructure+ADCPCheckpoint 1Checkpoint 168
SlocumCTD+ADCPFalklandsRDV60
Table 3: Fleet details for deployment around South Georgia and the Falklands
Research StationsAUV TypesNumber RequiredTask Duration (days)Earliest StartLatest End
S1Slocum - CTD+microstructure+ADCP272025-01-132025-08-09
S1Slocum - CTD+ADCP1102025-01-162025-01-27
S1Slocum - CTD+optode+PAR+Ecopuck182025-01-112025-01-31
S2Slocum - CTD+microstructure+ADCP272025-01-212025-03-07
S2ALR1500 - UVP6+EK80+EcoBBRTD1152025-01-112025-01-31
S2Slocum - CTD+optode+PAR+Ecopuck1102025-01-162025-01-31
Table 4: Requested science surveys around South Georgia and the Falklands
Figure 4: Timeline of MAV activities for the deployment around South Georgia and the Falklands
Figure 5: Heatmaps of completion rates of task requirements for the deployment around South Georgia and the Falklands: (a) with the research ship and mixed initial battery levels, and (b) without the research ship and with all MAVs deployed at 100% of battery
Figure 6: Case study integrating a ship with a pre-defined itinerary. The red dashed line with arrows represents the pre-defined route of the vessel travelling through waypoints shown by circles. The research stations are shown by the rectangles.
Research StationsAUV TypesNumber RequiredTask Duration (days)Earliest StartLatest End
S1Slocum - CTD+microstructure+ADCP1582025-01-032025-02-20
S1Slocum - CTD+ADCP10102025-01-162025-02-20
S1Slocum - CTD+optode+PAR+Ecopuck882025-01-112025-01-31
S2Slocum - CTD+microstructure+ADCP772025-01-132025-01-26
S2Slocum - CTD+ADCP1482025-01-182025-02-10
S2Slocum - CTD+optode+PAR+Ecopuck7102025-01-162025-01-31
Table 5: Requested science surveys around the United Kingdom
Figure 7: Time evolution of assets in plan. Panels (a) – (e). Each circle represents the asset location on the panel date, with the tails representing its path over the last 30 days, and colours indicating the asset type. The blue boxes represent the science locations and the red boxes the ship waypoint locations.
Figure 8: Top: worldwide overview of the seven Case Study 4 sites, lettered A-G and coloured consistently with the zoom panels; marker size is proportional to the number of MAVs deployed at each site, and sites sharing (near-)identical coordinates are spread apart slightly for legibility, connected to their true location by a thin leader line. Bottom: zoomed-in views of each site’s deployment/recovery locations (squares) and science survey locations (dots), with sites B-F (all in the North Atlantic/Europe) combined into a single panel.
DeploymentRegionSlocum GlidersALR1500Total MAVsScience TasksTotal Deployments
Walvis BayNamibia841236144
Appledore, CornwallCeltic Sea6410627
Cornwall to west coast of FranceEnglish Channel / Bay of Biscay10414638
REBELS-2 (Nuuk, Greenland)Labrador Sea155201261
BIO-Carbon ADF (Reykjavik, Iceland)North Atlantic154199298
eSWEETS3 (Aberdeen, Scotland)North Sea155201125
Table 6: Fleet, science task, and deployment summary for the seven Case Study 4 sites
Figure 9: Timeline of activities for the 59 distinct physical vehicles that make up the three fleets serving all seven Case Study 4 deployments (99 asset-deployments in total, with vehicles reused across consecutive projects within their fleet); colour indicates activity type, dark ”Logistics” bars mark the transit legs between consecutive deployments of the same fleet, and the thin overlay strip beneath each bar shades the asset’s battery level (see colour bar) over time.
Figure 10: Snapshot of the BIO-Carbon ADF (Reykjavik, Iceland) plan on 15 September 2024, roughly six weeks into the deployment, framed over the same North Atlantic/Europe extent as the combined zoom panel of Figure 8 . Coloured trails show each MAV’s PolarRoute-routed track over the preceding two weeks (Slocum gliders in green, ALR1500 vehicles in purple), with a solid marker at its position on the snapshot date; black squares mark the deployment/recovery location and black dots mark science survey locations.
Figure 11: Additional science days gained per instrument/sensor type under three hypothetical fleet upgrades (a 10% increase in battery capacity, a 10% increase in transit speed, and both combined), relative to the standard fleet configuration; radial axis on a log scale.
More in this topic

Multi-Robot Task Allocation

Multi-robot task allocation assigns work across robot teams while balancing capabilities, costs, timing, and coordination constraints.

See all 114 papers

Explore similar work

CardsList

  1. Fleet-Level Battery-Health-Aware Scheduling for Autonomous Mobile Robots

    Mar 24, 2026Jiachen Li, Shihao Li, Jian Chu +2Multi-Robot Task AllocationMulti-Agent System Optimization

    University of Texas at Austin, Austin, TX 78712 USA

  2. An Agentic AI Framework with Large Language Models and Chain-of-Thought for UAV-Assisted Logistics Scheduling with Mobile Edge Computing

    May 13, 2026Hanwen Zhang, Dusit Niyato, Wei Zhang +2Large Language Model-Guided OptimizationEdge Computing

    College of Computing and Data Science, Nanyang Technological University, Singapore, 639798, Singapore · Information and Communications Technology Cluster, Singapore Institute of Technology, Singapore, 828608, Singapore

  3. Joint Optimization of Trajectory Control, Resource Allocation, and Task Offloading for Multi-UAV-Assisted IoV

    May 6, 2026Maoxin Ji, Qiong Wu, Pingyi Fan +4Trajectory OptimizationIntelligent Transportation Systems

    School of Internet of Things Engineering, Jiangnan University, Wuxi 214122, China · School of Information Engineering, Jiangxi Provincial Key Laboratory of Advanced Signal Processing and Intelligent Communications, Nanchang University, Nanchang 330031, China · Department of Electronic Engineering, State Key laboratory of Space Network and Communications, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing 100084, China +4