Survey on autonomous surface vessels

Part 1 - A new detailed definition of autonomy levels

Conference Paper (2017)
Author(s)

Matteo Schiaretti (Student TU Delft)

Linying Chen (TU Delft - Transport Engineering and Logistics)

Rudy R. Negenborn (TU Delft - Transport Engineering and Logistics)

DOI related publication
https://doi.org/10.1007/978-3-319-68496-3_15 Final published version
More Info
expand_more
Publication Year
2017
Language
English
Pages (from-to)
219-233
Publisher
Springer
ISBN (print)
978-3-319-68495-6
Event
ICCL 2017: 8th International Conference on Computational Logistics (2017-10-18 - 2017-10-20), Southampton, United Kingdom
Downloads counter
171

Abstract

Autonomous Surface Vessels (ASVs) have been involved in numerous projects since the 1990s. Many ASV projects have been successfully realized, and as many are still under development. Together with the development of those new autonomous vessels, the research on classification about ASVs has become important. The classifications provide clarity to researchers, designers, shipbuilders, equipment manufacturers, ship owners and operators, enabling accurate specification of the desired level of autonomy in design and operations. Moreover, the involved research paves the way to a clearer understanding of the opportunity and challenges of research on autonomous vehicles. In this paper, we introduce the emerging concept of autonomous vessels. A multi-layer multi-agent control architecture of cooperative transport systems from the perspective of ASVs is proposed. Moreover, we provide an overview of existing research on the classification of autonomy. Based on the analysis, a detailed definition and categorization of autonomy levels for ASVs is proposed starting from the characteristics of ASVs and existing classification of autonomy. The proposed autonomy levels categorization assesses the overall autonomy level of a vessel by analyzing the automated sub-systems: Decision, Actions, Exceptions, and Cooperation. This categorization can be used to analyze existing ASV prototypes to gain insight into the status and trend of ASV research.