Appendix D

Levels of Automation

Unmanned systems can be automated or autonomous. Automated systems or equipment are controlled by a rigid set of rules that allow no deviations. This kind of automation assumes that an operator performs any requirements and decision making that the automated sequence is not instructed to do through its set of rigid rules.¹ In contrast, autonomous systems can independently formulate multiple courses of action and can select one to pursue in order to achieve the objectives prescribed to it. In essence, autonomous systems imply a higher order of autonomy, and their development requires the introduction of multiple automation sequences and artificial intelligence to enable equipment to work autonomously without the need of an operator.

“Level of automation” refers to the degree to which a task is automated; it reflects the level of human control versus computer control of a given task. Over the years equipment and vehicle developers and manufacturers have established a set of defining levels of automation to fit their industries. However, because of the unique operating environment in the maritime domain, autonomy levels or degrees used in other modes of transport, such as the automotive industry, are not considered to be directly transferrable to, for example, vessels.

Efforts to standardize terms applicable to the development and evaluation of unmanned systems for maritime domain awareness (MDA) need to account for an array of factors such as level of human interaction, task

__________________

¹ Defense Science Board. 2016. “Summer Study on Autonomy.” https://fas.org/irp/agency/dod/dsb/autonomy-ss.pdf.

complexity, and operational environment.² The degree to which these factors apply to a given system provide valuable information about not only the developers and analysts, but also the end user. To that end, this section offers three examples of competing options for evaluating the various levels or degrees of autonomy under which a vessel may operate. These standards are illustrative only and do not represent all the various proposals in this area. Essentially, the international debate about what may emerge as an internationally recognized standard is ongoing.

NATIONAL INSTITUTE OF STANDARDS AND TECHNOLOGY

The National Institute of Standards and Technology (NIST) in collaboration with other government and industry experts defined four levels of automation in its Autonomy Levels for Unmanned System Framework.³ First, NIST’s framework defines the term autonomous operations of an unmanned system (UMS) as those in which the UMS “receives its mission from either the operator who is off the UMS or another system that the UMS interacts with and accomplishes that mission with or without further human-robot interaction.”⁴

The framework also defines four levels of UMSs. These levels are designed to offer a structure of increasing autonomy, with automation rising from one level to the next.

Remote Control

“A mode of UMS operation wherein the human operator controls the UMS on a continuous basis, from a location off the UMS via only her/his direct observation. In this mode, the UMS takes no initiative and relies on continuous or nearly continuous input from the human operator.”⁵

__________________

² H. Huang, J. Albus, E. Messina, R. Wade, and W. English. 2004. “Specifying Autonomy Levels for Unmanned Systems: Interim Report.” SPIE Defense and Security Symposium 2004, Conference 5422, Orlando, Florida. https://www.researchgate.net/publication/228767910_Specifying_autonomy_levels_for_unmanned_systems_interim_report.

³ National Institute of Standards and Technology. 2008. “Autonomy Levels for Unmanned Systems (ALFUS) Framework, Volume I: Terminology.” NIST Special Report 1011-I-2.0. https://www.nist.gov/sites/default/files/documents/el/isd/ks/NISTSP_1011-I-2-0.pdf.

⁴ National Institute of Standards and Technology. 2008. “Autonomy Levels for Unmanned Systems (ALFUS) Framework, Volume I: Terminology.” NIST Special Report 1011-I-2.0, p. 15. https://www.nist.gov/sites/default/files/documents/el/isd/ks/NISTSP_1011-I-2-0.pdf.

⁵ National Institute of Standards and Technology. 2008. “Autonomy Levels for Unmanned Systems (ALFUS) Framework, Volume I: Terminology.” NIST Special Report 1011-I-2.0, p. 23. https://www.nist.gov/sites/default/files/documents/el/isd/ks/NISTSP_1011-I-2-0.pdf.

Teleoperation

“A mode of UMS operation wherein the human operator, using sensory feedback, either directly controls the actuators or assigns incremental goals on a continuous basis, from a location off the UMS.”⁶

Semi-Autonomous

“A mode of UMS operation wherein the human operator and/or the UMS plan(s) and conduct(s) a mission and requires various levels of human-robot interaction. The UMS is capable of autonomous operation in between the human interactions.”⁷

Fully Autonomous

“A mode of UMS operation wherein the UMS accomplishes its assigned mission, within a defined scope, without human intervention while adapting to operational and environmental conditions.”⁸

INTERNATIONAL MARITIME ORGANIZATION

Another framework, specifically for unmanned surface ships, has also been developed by the International Maritime Organization (IMO)—the global regulatory body for international shipping.⁹ The IMO defines a Maritime Autonomous Surface Ship (MASS) as a ship with decision support system(s) that, to various degrees, can operate independently of human interactions. The IMO defined four degrees of autonomy for MASS. The framework recognizes, however, that MASS could operate at one or more of these levels within the duration of a voyage. The four degrees are as follows, although a MASS could operate at one or more degrees of autonomy during the duration of a single voyage:

__________________

⁶ National Institute of Standards and Technology. 2008. “Autonomy Levels for Unmanned Systems (ALFUS) Framework, Volume I: Terminology.” NIST Special Report 1011-I-2.0, p. 23. https://www.nist.gov/sites/default/files/documents/el/isd/ks/NISTSP_1011-I-2-0.pdf.

⁷ National Institute of Standards and Technology. 2008. “Autonomy Levels for Unmanned Systems (ALFUS) Framework, Volume I: Terminology.” NIST Special Report 1011-I-2.0, p. 23. https://www.nist.gov/sites/default/files/documents/el/isd/ks/NISTSP_1011-I-2-0.pdf.

⁸ National Institute of Standards and Technology. 2008. “Autonomy Levels for Unmanned Systems (ALFUS) Framework, Volume I: Terminology.” NIST Special Report 1011-I-2.0, p. 22. https://www.nist.gov/sites/default/files/documents/el/isd/ks/NISTSP_1011-I-2-0.pdf.

⁹ International Maritime Organization. 2018. “IMO takes first steps to address autonomous ships.” http://www.imo.org/en/MediaCentre/PressBriefings/Pages/08-MSC-99-MASS-scoping.aspx.

1. “Ship with automated processes and decision support: Seafarers are on board to operate and control shipboard systems and functions. Some operations may be automated.
2. Remotely controlled ship with seafarers on board: The ship is controlled and operated from another location, but seafarers are on board.
3. Remotely controlled ship without seafarers on board: The ship is controlled and operated from another location. There are no seafarers on board.
4. Fully autonomous ship: The operating system of the ship is able to make decisions and determine actions by itself.”¹⁰

The IMO’s aim was to complete the scoping exercise by 2020, and the next steps in the IMO’s agenda to scope MASS were set to (1) identify current provisions of IMO instruments and assess whether and how they could be applicable to ships with varying degrees of autonomy and (2) conduct an analysis to determine the most appropriate way to address MASS operations, taking into account the human interactions, state of technology, and operational factors. This work is progressing rapidly; the first step has been completed, and in June 2019, the IMO approved a set of interim guidelines for MASS trials.¹¹

The next step is to analyze and determine the most appropriate way to address MASS operations, taking into account, inter alia, human element, technology, and operational factors to identify the need for:

• Equivalences as provided for by the instruments or developing interpretations; and/or
• Amending existing instruments; and/or
• Developing new instruments; or
• None of the above as a result of the analysis.

Reporting of the updated results of this regulatory scoping exercise for the use of MASS was included on the agenda for MSC 102.¹²

__________________

¹⁰ International Maritime Organization. 2018. http://www.imo.org/en/MediaCentre/PressBriefings/Pages/08-MSC-99-MASS-scoping.aspx.

¹¹ World Maritime News. 2019. “IMO Approves Autonomous Ship Trial Guidelines.” https://worldmaritimenews.com/archives/279047/imo-approves-autonomous-ship-trial-guidelines.

¹² MSC 102 was scheduled for May 13–22, 2020, but because of the impact of COVID-19, was postponed without a new date determined.

TABLE D-1 Levels of Control from the UK Marine Industry Alliance

UK MARINE INDUSTRIES ALLIANCE

In addition, the UK Marine Industries Alliance Code of Practice for Maritime Autonomous Systems (Surface) for craft less than 24 meters (version 3.0, November 2019)¹³ recognizes six levels of control: Manned, Operated, Directed, Delegated, Monitored, and Autonomous. Table D-1, from the Maritime Autonomous Surface Ships Industry Conduct Principles and Code of Practice, describes these six levels.

__________________

¹³ Maritime UK. 2019. “Maritime Autonomous Surface Ships (MASS) UK Industry Conduct Principles and Code of Practice.” https://www.maritimeuk.org/media-centre/publications/maritime-autonomous-surface-ships-industry-conduct-principles-code-practice.

As stated in this UK Code of Practice, “these definitions for Level of Control should be considered alongside the IMO Degrees of Autonomy since levels of control may be different for different functions aboard the same MASS.”¹⁴

BUREAU VERITAS

Also within the context of maritime shipping vessels, in October 2019 the international certification agency Bureau Veritas published its Guidelines for Autonomous Shipping. In this guidance, Bureau Veritas defines the following levels of automation:¹⁵

• Degree A0—Human operated
• Degree A1—Human directed
• Degree A2—Human delegated
• Degree A3—Human supervised
• Degree A4—Full automation

Table D-2 from the Bureau Veritas guidance publication provides a quick description of the nature of these levels of shipping automation.

Finally, a fourth consideration may emerge from the International Organization for Standardization (ISO) because it has established a working group on Smart Shipping and a task group on MASS and is currently drafting new guidance on “Ships and marine technology—Terminology related to automation of Maritime Autonomous Surface Ships (MASS).”¹⁶ It is too soon to evaluate the ISO’s proposed degrees of automation, but it will be worth monitoring.

NORWEGIAN FORUM FOR AUTONOMOUS SHIPS

The Norwegian Forum for Autonomous Ships considers “Operational autonomy levels” and bridge manning levels as presented in Table D-3.¹⁷

__________________

¹⁴ See UK Code of Practice, p. 21.

¹⁵ Bureau Veritas. 2019. “Guidance Note NI 641 DT R01 E: Guidelines for Autonomous Shipping.” Paris La Defense. http://erules.veristar.com/dy/data/bv/pdf/641-NI_2019-10.pdf.

¹⁶ Proposed terminology for MASS submitted by ISO, MSC 102/5/18 (February 11, 2020). http://nfas.autonomous-ship.org/resources/msc-102-5-18.pdf.

¹⁷ Combining the manning levels with the operational autonomy levels, this table defines different types of ship autonomy. http://nfas.autonomous-ship.org/resources/autonom-defs.pdf.

TABLE D-2 Degrees of Shipping Automation as Advised by Bureau Veritas

TABLE D-3 Operational Autonomy Levels from the Norwegian Forum for Autonomous Ships

AMERICAN BUREAU OF SHIPPING

The American Bureau of Shipping, Advisory on Autonomous Functionality¹⁸ identifies four main levels of automation based on “Sheridan’s levels of automation.” In increasing level of autonomy, the four levels from the American Bureau of Shipping specify the following:

1. Manual
   • No machine augmentation of human function
     • The computer offers no assistance: human must make all decisions and take all actions
2. Smart
   • Passive decision support: Human augmentation of human function
     • The computer offers a complete set of decisions/actions alternatives
     • The computer narrows the selection of decisions or actions to a few alternatives
     • The computer suggest one alternative
     • The computer executes the suggestion if the human approves
3. Semi-Autonomy
   • Human in the loop: human augmentation of machine function
     • The computer allows the human a restricted amount of time to veto the suggestion and then executes
     • The computer executes automatically, then informs the human

__________________

¹⁸ See https://ww2.eagle.org/en/innovation-and-technology/digital/autonomy/autonomyadvisory.html.

4. Autonomy
   • No human augmentation of machine function
     • The computer informs the human, but only if queried by the human
     • The computer informs the human, but only if the computer decides to do so
     • The computer makes all decisions, acts autonomously, and ignores the human

SOCIETY OF AUTOMOTIVE ENGINEERS

The Society of Automotive Engineers defines six levels of automotive automation,¹⁹ summarized in Table D-4.

INTERNATIONAL ORGANIZATION FOR STANDARDIZATION

Finally, a consideration may emerge from the ISO because it has established a working group on Smart Shipping and task group on MASS and are currently drafting new guidance on “Ships and marine technology—Terminology related to automation of Maritime Autonomous Surface Ships (MASS).”²⁰ One proposal from the ISO to the IMO is to define automation as “pertaining to a process or device that, under specified conditions, can function without human intervention (definition is based on ISO/TR 11065).” It is too soon to evaluate the ISO’s proposed degrees of automation, but it will be worth monitoring.

__________________

¹⁹ Intel Explainer: 6 Levels of Autonomous Driving. https://newsroom.intel.com/news/autonomous-driving-hands-wheel-no-wheel-all/#gs.9pb4gk.

²⁰ ISO/AWI 23860. Ships and marine technology—Terminology related to automation of Maritime Autonomous Surface Ships (MASS). https://www.iso.org/standard/77186.html; see also proposed terminology for MASS Submitted by ISO, MSC 102/5/18 (February 11, 2020). http://nfas.autonomous-ship.org/resources/msc-102-5-18.pdf.

TABLE D-4 Levels of Automotive Automation from the Society of Automotive Engineers