MarineLLM-PDDL: Generation of Planning Domains for Marine Vessels Using Past Incident Response Plans

Mahya Mohammadi Kashani; Stefan Heinrich; Andrzej Wasowski

doi:10.1007/978-3-031-89471-8_47

MarineLLM-PDDL: Generation of Planning Domains for Marine Vessels Using Past Incident Response Plans

Mahya Mohammadi Kashani
, Stefan Heinrich
, Andrzej Wasowski

Research output: Conference Article in Proceeding or Book/Report chapter › Article in proceedings › Research › peer-review

Abstract

Testing the hardware and software of marine vessels in field trials is a necessity to avoid technical and environmental catastrophes. Conducting tests with large vessels is costly. Multiple realistic domain descriptions based on past missions could increase the value of simulation tests, reducing the need for expensive field tests. In this paper, we generate scenarios from unstructured Incident Response Plan (IRP) documents using Large Language Models (LLMs), converting them to standard structured planning programs. The two synthesized marine test-domain datasets contain approximately 90% parsable, 75% solvable, and 57% correct planning programs.

Original language	English
Title of host publication	Springer Proceedings in Advanced Robotics
Number of pages	7
Volume	36
Place of Publication	Springer Cham
Publisher	Springer
Publication date	2025
Pages	307-313
ISBN (Print)	978-3-031-89470-1
ISBN (Electronic)	978-3-031-89471-8
DOIs	https://doi.org/10.1007/978-3-031-89471-8_47
Publication status	Published - 2025
Event	Conference on European Robotics Forum - Kultur- & Kongresszentrum Liederhalle, Stuttgart, Germany Duration: 25 Mar 2025 → 27 Mar 2025 https://erf2025.eu/?utm

Conference

Conference	Conference on European Robotics Forum
Location	Kultur- & Kongresszentrum Liederhalle
Country/Territory	Germany
City	Stuttgart
Period	25/03/2025 → 27/03/2025
Internet address	https://erf2025.eu/?utm

Series	European Robotics Forum

Keywords

Mission Planning
Marine Vessels
Large Language Models (LLMs)
Scenarios Generation
Test Domain

Access to Document

10.1007/978-3-031-89471-8_47

Cite this

@inproceedings{78ed94dd58f34e94a723b8207de56878,

title = "MarineLLM-PDDL: Generation of Planning Domains for Marine Vessels Using Past Incident Response Plans",

abstract = "Testing the hardware and software of marine vessels in field trials is a necessity to avoid technical and environmental catastrophes. Conducting tests with large vessels is costly. Multiple realistic domain descriptions based on past missions could increase the value of simulation tests, reducing the need for expensive field tests. In this paper, we generate scenarios from unstructured Incident Response Plan (IRP) documents using Large Language Models (LLMs), converting them to standard structured planning programs. The two synthesized marine test-domain datasets contain approximately 90\% parsable, 75\% solvable, and 57\% correct planning programs.",

keywords = "Mission Planning, Marine Vessels, Large Language Models (LLMs), Scenarios Generation, Test Domain",

author = "\{Mohammadi Kashani\}, Mahya and Stefan Heinrich and Andrzej Wasowski",

year = "2025",

doi = "10.1007/978-3-031-89471-8\_47",

language = "English",

isbn = "978-3-031-89470-1",

volume = "36",

series = "European Robotics Forum",

publisher = "Springer",

pages = "307--313",

booktitle = "Springer Proceedings in Advanced Robotics",

address = "Germany",

note = "Conference on European Robotics Forum, ERF ; Conference date: 25-03-2025 Through 27-03-2025",

url = "https://erf2025.eu/?utm",

}

Mohammadi Kashani, M, Heinrich, S & Wasowski, A 2025, MarineLLM-PDDL: Generation of Planning Domains for Marine Vessels Using Past Incident Response Plans. in Springer Proceedings in Advanced Robotics . vol. 36, Springer, Springer Cham, European Robotics Forum, pp. 307-313, Conference on European Robotics Forum, Stuttgart, Baden-Württemberg, Germany, 25/03/2025. https://doi.org/10.1007/978-3-031-89471-8_47

MarineLLM-PDDL: Generation of Planning Domains for Marine Vessels Using Past Incident Response Plans. / Mohammadi Kashani, Mahya; Heinrich, Stefan ; Wasowski, Andrzej.
Springer Proceedings in Advanced Robotics . Vol. 36 Springer Cham: Springer, 2025. p. 307-313 (European Robotics Forum).

Research output: Conference Article in Proceeding or Book/Report chapter › Article in proceedings › Research › peer-review

TY - GEN

T1 - MarineLLM-PDDL: Generation of Planning Domains for Marine Vessels Using Past Incident Response Plans

AU - Mohammadi Kashani, Mahya

AU - Heinrich, Stefan

AU - Wasowski, Andrzej

PY - 2025

Y1 - 2025

N2 - Testing the hardware and software of marine vessels in field trials is a necessity to avoid technical and environmental catastrophes. Conducting tests with large vessels is costly. Multiple realistic domain descriptions based on past missions could increase the value of simulation tests, reducing the need for expensive field tests. In this paper, we generate scenarios from unstructured Incident Response Plan (IRP) documents using Large Language Models (LLMs), converting them to standard structured planning programs. The two synthesized marine test-domain datasets contain approximately 90% parsable, 75% solvable, and 57% correct planning programs.

AB - Testing the hardware and software of marine vessels in field trials is a necessity to avoid technical and environmental catastrophes. Conducting tests with large vessels is costly. Multiple realistic domain descriptions based on past missions could increase the value of simulation tests, reducing the need for expensive field tests. In this paper, we generate scenarios from unstructured Incident Response Plan (IRP) documents using Large Language Models (LLMs), converting them to standard structured planning programs. The two synthesized marine test-domain datasets contain approximately 90% parsable, 75% solvable, and 57% correct planning programs.

KW - Mission Planning

KW - Marine Vessels

KW - Large Language Models (LLMs)

KW - Scenarios Generation

KW - Test Domain

U2 - 10.1007/978-3-031-89471-8_47

DO - 10.1007/978-3-031-89471-8_47

M3 - Article in proceedings

SN - 978-3-031-89470-1

VL - 36

T3 - European Robotics Forum

SP - 307

EP - 313

BT - Springer Proceedings in Advanced Robotics

PB - Springer

CY - Springer Cham

T2 - Conference on European Robotics Forum

Y2 - 25 March 2025 through 27 March 2025

ER -

MarineLLM-PDDL: Generation of Planning Domains for Marine Vessels Using Past Incident Response Plans

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this