195
Improving Safety
the International Network for Autonomous Ships ( INAS, 2020) as an extension of
NFAS outside Norway. The research centre for Autonomous Marine Operations and
Systems ( AMOS, 2020) at NTNU was established in 2013 as a multidisciplinary
centre for autonomous marine operations and control systems.
More extensive research projects, such as AAWA ( 2020), MUNIN ( 2020), Autosea
( 2020), Autoship ( 2020) and IMAT ( 2020), focus on specific concepts where unmanned,
autonomous or smart ships are explored and tested. The world’s first fully electric and
autonomous container ship, Yara Birkeland ( 2020), is under construction. The ship is
now planned to be in operation by 2022, earlier planned to start in 2020, and centres are
scheduled to handle all aspects of remote and autonomous operation to ensure safety.
A newly established company, Zeabuz ( 2020), will test prototypes of an autonomous electric ferry system for urban waterways. Limited information is given about
the concept other than it will be s elf-driving and electric. The remote and autonomous operational aspect of an RCC is not mentioned, but a remote support center is
planned to operate in the initial phase.
Most of the projects above are in the initial stages with limited operational experience. Most safety concerns are related to the reliability of sensors and technical
equipment and their ability to handle different situations.
Experiences Related to Safety Challenges
In operation, MASS have only been tested in small scale without an interface for human
supervision or control. We have examples of safety issues during early testing of autonomous technology ( software and hardware) local in Norway in Trondheimsfjorden, with
the small-scale version of the passenger ferry AutoFerry. One example is loss of control
due to a technical failure, a s o-called fallout, of the dynamic positioning system which
made AutoFerry run into the harbour. However, there is no systematic data collection
of failures or unforeseen events, and this is not a requirement from the Norwegian
Maritime Authority ( NMA) at present. Though, a Preliminary Hazard Analysis ( PHA)
has been carried out for the operation of the AutoFerry ( Thieme et al., 2019), the main
hazards were software failure; failure of internal and external communication systems;
traffic in the channel ( especially kayaks, difficult to discover); passenger handling and
monitoring; and weather conditions. The practical challenges encountered in the ferry
project were also listed. These challenges are related to available risk analysis methods
and data, determining and establishing an equivalent safety level, and some of the prescriptive regulations currently in use by NMA. At present ( start 2021) the AutoFerry
project lacks an established plan on who should operate the ferry and how to intervene
especially during emergencies. The human operator is said to be in the loop and able
to intervene from an RCC. However, none of the projects have developed such a centre
or made detailed plans for their operation so far. In the reviewed projects, the focus has
been on technology development.
A literature review on risk identification methods for MASS ( Hoem, 2019) identifies the uncertainty of the operational mode and context of the MASS operation ( i.e.
operational domain) to be a major challenge when identifying operational hazards
and risks. There is a need to define what conditions the ship is designed to operate
under. Rødseth ( 2018) proposed to use the “ operational design domain” from SAE
J3016 ( 2018) to define the context, i.e. the operational domain with its complexity.
Improving Safety
the International Network for Autonomous Ships ( INAS, 2020) as an extension of
NFAS outside Norway. The research centre for Autonomous Marine Operations and
Systems ( AMOS, 2020) at NTNU was established in 2013 as a multidisciplinary
centre for autonomous marine operations and control systems.
More extensive research projects, such as AAWA ( 2020), MUNIN ( 2020), Autosea
( 2020), Autoship ( 2020) and IMAT ( 2020), focus on specific concepts where unmanned,
autonomous or smart ships are explored and tested. The world’s first fully electric and
autonomous container ship, Yara Birkeland ( 2020), is under construction. The ship is
now planned to be in operation by 2022, earlier planned to start in 2020, and centres are
scheduled to handle all aspects of remote and autonomous operation to ensure safety.
A newly established company, Zeabuz ( 2020), will test prototypes of an autonomous electric ferry system for urban waterways. Limited information is given about
the concept other than it will be s elf-driving and electric. The remote and autonomous operational aspect of an RCC is not mentioned, but a remote support center is
planned to operate in the initial phase.
Most of the projects above are in the initial stages with limited operational experience. Most safety concerns are related to the reliability of sensors and technical
equipment and their ability to handle different situations.
Experiences Related to Safety Challenges
In operation, MASS have only been tested in small scale without an interface for human
supervision or control. We have examples of safety issues during early testing of autonomous technology ( software and hardware) local in Norway in Trondheimsfjorden, with
the small-scale version of the passenger ferry AutoFerry. One example is loss of control
due to a technical failure, a s o-called fallout, of the dynamic positioning system which
made AutoFerry run into the harbour. However, there is no systematic data collection
of failures or unforeseen events, and this is not a requirement from the Norwegian
Maritime Authority ( NMA) at present. Though, a Preliminary Hazard Analysis ( PHA)
has been carried out for the operation of the AutoFerry ( Thieme et al., 2019), the main
hazards were software failure; failure of internal and external communication systems;
traffic in the channel ( especially kayaks, difficult to discover); passenger handling and
monitoring; and weather conditions. The practical challenges encountered in the ferry
project were also listed. These challenges are related to available risk analysis methods
and data, determining and establishing an equivalent safety level, and some of the prescriptive regulations currently in use by NMA. At present ( start 2021) the AutoFerry
project lacks an established plan on who should operate the ferry and how to intervene
especially during emergencies. The human operator is said to be in the loop and able
to intervene from an RCC. However, none of the projects have developed such a centre
or made detailed plans for their operation so far. In the reviewed projects, the focus has
been on technology development.
A literature review on risk identification methods for MASS ( Hoem, 2019) identifies the uncertainty of the operational mode and context of the MASS operation ( i.e.
operational domain) to be a major challenge when identifying operational hazards
and risks. There is a need to define what conditions the ship is designed to operate
under. Rødseth ( 2018) proposed to use the “ operational design domain” from SAE
J3016 ( 2018) to define the context, i.e. the operational domain with its complexity.