AI for Autonomous Ships: Challenges in Design and Validation

Research output: Contribution to conferenceConference AbstractResearch

Abstract

The development of several key elements in autonomous shipping, such as situational awareness systems and autonomous navigational decision-making, are likely to be based on various artificial intelligence (AI) technologies. The emergence of these new technologies will bring along new challenges for all stakeholders involved in their development and approval. Several key challenges in different development phases will be presented with selected new approaches proposed to support the development activities. Machine learning technologies, for example, impose challenges regarding data collection, transfer and processing. The design, training and validation, as well as transparency of such systems raise further questions. In validation, the lack of prescriptive standards further complicates the design and approval process. Additionally, when in operation, the AI technologies may introduce new kinds of risks and security vulnerabilities.
LanguageEnglish
Publication statusPublished - 21 Mar 2018
EventInternational Seminar on Safety and Security of Autonomous Vessels, ISSAV 2018 - TU Delft, Delft, Netherlands
Duration: 21 Mar 201821 Mar 2018

Seminar

SeminarInternational Seminar on Safety and Security of Autonomous Vessels, ISSAV 2018
Abbreviated titleISSAV 2018
CountryNetherlands
CityDelft
Period21/03/1821/03/18

Fingerprint

Artificial intelligence
Ships
Freight transportation
Transparency
Learning systems
Decision making
Processing

Keywords

  • autonomous ship
  • artificial intelligence
  • validation

Cite this

Heikkilä, E. (2018). AI for Autonomous Ships: Challenges in Design and Validation. Abstract from International Seminar on Safety and Security of Autonomous Vessels, ISSAV 2018, Delft, Netherlands.
Heikkilä, Eetu. / AI for Autonomous Ships : Challenges in Design and Validation. Abstract from International Seminar on Safety and Security of Autonomous Vessels, ISSAV 2018, Delft, Netherlands.
@conference{e4738c832b4c4e26bc24649e54935a89,
title = "AI for Autonomous Ships: Challenges in Design and Validation",
abstract = "The development of several key elements in autonomous shipping, such as situational awareness systems and autonomous navigational decision-making, are likely to be based on various artificial intelligence (AI) technologies. The emergence of these new technologies will bring along new challenges for all stakeholders involved in their development and approval. Several key challenges in different development phases will be presented with selected new approaches proposed to support the development activities. Machine learning technologies, for example, impose challenges regarding data collection, transfer and processing. The design, training and validation, as well as transparency of such systems raise further questions. In validation, the lack of prescriptive standards further complicates the design and approval process. Additionally, when in operation, the AI technologies may introduce new kinds of risks and security vulnerabilities.",
keywords = "autonomous ship, artificial intelligence, validation",
author = "Eetu Heikkil{\"a}",
year = "2018",
month = "3",
day = "21",
language = "English",
note = "International Seminar on Safety and Security of Autonomous Vessels, ISSAV 2018, ISSAV 2018 ; Conference date: 21-03-2018 Through 21-03-2018",

}

Heikkilä, E 2018, 'AI for Autonomous Ships: Challenges in Design and Validation' International Seminar on Safety and Security of Autonomous Vessels, ISSAV 2018, Delft, Netherlands, 21/03/18 - 21/03/18, .

AI for Autonomous Ships : Challenges in Design and Validation. / Heikkilä, Eetu.

2018. Abstract from International Seminar on Safety and Security of Autonomous Vessels, ISSAV 2018, Delft, Netherlands.

Research output: Contribution to conferenceConference AbstractResearch

TY - CONF

T1 - AI for Autonomous Ships

T2 - Challenges in Design and Validation

AU - Heikkilä, Eetu

PY - 2018/3/21

Y1 - 2018/3/21

N2 - The development of several key elements in autonomous shipping, such as situational awareness systems and autonomous navigational decision-making, are likely to be based on various artificial intelligence (AI) technologies. The emergence of these new technologies will bring along new challenges for all stakeholders involved in their development and approval. Several key challenges in different development phases will be presented with selected new approaches proposed to support the development activities. Machine learning technologies, for example, impose challenges regarding data collection, transfer and processing. The design, training and validation, as well as transparency of such systems raise further questions. In validation, the lack of prescriptive standards further complicates the design and approval process. Additionally, when in operation, the AI technologies may introduce new kinds of risks and security vulnerabilities.

AB - The development of several key elements in autonomous shipping, such as situational awareness systems and autonomous navigational decision-making, are likely to be based on various artificial intelligence (AI) technologies. The emergence of these new technologies will bring along new challenges for all stakeholders involved in their development and approval. Several key challenges in different development phases will be presented with selected new approaches proposed to support the development activities. Machine learning technologies, for example, impose challenges regarding data collection, transfer and processing. The design, training and validation, as well as transparency of such systems raise further questions. In validation, the lack of prescriptive standards further complicates the design and approval process. Additionally, when in operation, the AI technologies may introduce new kinds of risks and security vulnerabilities.

KW - autonomous ship

KW - artificial intelligence

KW - validation

UR - https://d1rkab7tlqy5f1.cloudfront.net/Websections/Delft%20Safety%20and%20Security/ISSAV2018%20%28P7%29_Heikkila.pdf

M3 - Conference Abstract

ER -

Heikkilä E. AI for Autonomous Ships: Challenges in Design and Validation. 2018. Abstract from International Seminar on Safety and Security of Autonomous Vessels, ISSAV 2018, Delft, Netherlands.