Demonstrating a strong system safety argument is a critical part of any safety assurance process.
This demonstration attests to an autonomous system’s capacity to meet safety requirements and is a crucial step for enabling true integration of highly automated and autonomous systems into the air, maritime, and land domains. Without this process it is difficult to demonstrate any sort of confidence in an autonomous system’s safety and reliability characteristics.
However, current system safety processes assume a level of human control and authority during operation as well as other assumptions around the level of autonomy, acceptable mitigations etc. of a system. This makes it difficult for more highly automated and autonomous systems to demonstrate complete compliance with these safety analyses and their requirements.
TAS is proud to announce that we are working with Frazer-Nash Consultancy to develop a safety assurance framework to help tackle these challenges as a part of the TAS Body of Knowledge.
The framework is intended to augment, not replace, the current system safety processes and provide complementary techniques, processes, and methods relevant to highly automated and autonomous systems. This will create the stepping-stones needed to bridge those gaps between the current processes and what is needed to truly integrate these systems into the domains.
The project will culminate in high level guidance documentation on the overall system safety analysis process. It will also produce a repository of useful standards and research that will assist in completing detailed safety analyses referenced in this documentation.
If you are interested in finding out more or would like more information on the TAS Body of Knowledge, please contact us at NASFP@tasdcrc.com.au.
By Beth Cardier, TAS Fellow
How can we know whether a transfer of meaning between a person and a machine is meaningful enough for control which is morally responsible?
This question was posed by NATO’s Science and Technology Organisation in a workshop run in collaboration with the Netherlands Organisation for Applied Scientific Research (TNO) and Trusted Autonomous Systems (TAS). The topic was Meaningful Human Control of AI-based Systems – Key Characteristics, Influencing Factors and Design Considerations (further details are below.) On this panel, Dr Kate Devitt and myself spoke on the complexity of determining meaning in an open world.
From my perspective as a narrative analyst, the question revolves around the transfer of information between humans and machines. When humans communicate with each other, one person must successfully transfer what’s in their head to another person’s mind in order for correct interpretation to be possible. Humans also draw on surrounding contexts for accurate interpretation, using an awareness of past, future, and theory of mind to receive the correct meaning. When humans transfer information or instructions to machines, however, machines are not able to process these complexities, so key information can be lost.
There is also a problem of asynchronicity in human-machine teams, unless the machine can genuinely adapt. During the machine’s construction, engineers created the channels of mean-production it would use long before it even met its human co-worker. The human has to compensate for this lag, compensating for a machine that relies on pre-determined meaning structures and other, out-of-sight contexts related to its design.
An ordinary user understands this disconnect, even though they might not consciously analyse it. It is why machines still receive aghast reactions on social media when they try to emulate humans, regardless of how amusing or charming their presentations.
This is one of the reasons I focus on adaptive communication in my research. My goal is to introduce a new suite of information structures to human-machine information exchange so that our technologies can be more flexible. When humans exchange information with each other, narrative emerges because the language tokens we use are inadequate. Stories form bridging structure between you and I, or between past and present, and are the means of bringing another person along when our circumstances exceed an initial plan. Science has not fully exploited these adaptive information structures yet.
For example, consider these transformations. Information can have properties of water, taking the shape of the vessel it fills, that vessel being context. Or information can have properties of a vine, reaching beyond itself to grow from one place to the next. Or maybe information is a knife, dividing the swim of experience into discrete objects, and then cutting from such a different angle that reinterpretation is required. We want our machines to act reliably in the open world but the world itself is not reliable. How can we maintain meaningful communication with machines, when accurate interpretation requires a vine or a knife or a river?
At a Glance
- Workshop: NATO’s Science and Technology Organisation ran a workshop in collaboration with the Netherlands Organisation for Applied Scientific Research (TNO) and Trusted Autonomous Systems (TAS).
- Focus: Meaningful Human Control of AI-based Systems – Key Characteristics, Influencing Factors and Design Considerations
- When: October 27, 2021
- Where: Berlin and online.
- Speakers: Dr Daniele Amoroso of the International Committee for Robot Arms Control, Dr Le on Kester of TNO, Dr Luciano Cavalcante Siebert assistant professor at the Interactive Intelligence Group at Delft University of Technology, Dr Kate Devitt from TAS and Dr Beth Cardier, from Griffith University and TAS.
Image: Future Memory, Oil on linen, 122 x 137 cm, 2021 by Kathryn Brimblecombe-Fox
By Rachel Horne, Assurance of Autonomy Activity Lead, TAS
“Autonomous and Remotely Operated Vessels: 2021 to 2040”  is a newly published paper considering the future of autonomous vessels. It forms part of the Maritime Industry Australia Limited (MIAL) Future Leaders White Paper: Predictions for the Australian Maritime Industry 2040.
The paper considers: What are autonomous vessels? How are they regulated? Why is regulation so difficult and what can we change to make it easier? A central focus for the discussion is how we might accelerate development of autonomous vessels with a focus on building sovereign capability and what success in this sector would mean for the Australian autonomous systems ecosystem in 2040.
The paper aims to introduce autonomous vessels and the opportunities and challenges the technology presents for Australia, in a way that is accessible and thought-provoking. It explores a broad range of topics and issues related to autonomous vessels, starting with background on what these vessels are and why they are used, what their benefits are, and the features we will likely see by 2040 commercially and within Defence, and then moving on to impact on workforce, the current regulatory framework and conceptually difficult areas, and concluding with a range of proposed solutions.
Predictions for the Australian maritime industry in 2040
The paper includes six predictions for the Australian maritime industry in 2040, extracted below:
- Minimally crewed vessels with a spectrum of autonomous capabilities will be a normalised part of the commercial vessel fleet operating for routine passenger transport, movement of goods, scientific research, and tourism. The police, border protection and Defence agencies will have significant numbers of semi-autonomous and autonomous vessels in their fleets.
- A Maritime Water Space Management System (MWSPS) will have been implemented to manage allocation of surface and subsurface water space and interaction between smart vessels.
- The deconfliction service offered through the MWSPS, together with advanced navigation, sensing, and inter-vessel communication technologies, will enable minimal-crewing, and multiple semi-or fully-autonomous vessels to be supervised remotely by single operators, due to the significant reduction in collision risk.
- A new Commonwealth Government entity, ‘Australian Complex Autonomous Systems Safety Authority’ (ACASSA) will set the standards and conduct assurance activities for the “black box” behind autonomous and semi-autonomous systems, for each of the air, land and maritime domains. Two way secondments between ACASSA and traditional regulators will ensure a seamless experience for stakeholders, consistent regulatory and policy development, and the upskilling of staff.
- “RegTech” concepts will be implemented by the ACASSA to enable continuous background monitoring of AI-based autonomous systems, using risk thresholds to determine input required by the operator, and enabling non-intrusive compliance checks.
- Australian Ports are able to accommodate large international trading vessels with advanced autonomy on board, and the integration of Vessel Traffic Services with the Maritime Water Space Management System have reduced the workload of VTS operators and vessel crew, reduced incidents, and improved efficiency. 
How can the 2040 vision be achieved?
The paper argues that change, led by Government and supported by industry, is needed to ensure the maritime industry can access the range of benefits autonomous technology offers into the future. Collaboration, and a focus on new regulatory approaches, upskilling the maritime workforce, and smart ports, are all central to achieving the 2040 vision.
The paper concludes:
Autonomous and remotely operated vessels are already in operation in Australia and around the world, and their capability and availability are rapidly growing. It is predicted that, by 2040, these vessels will be an integrated, integral part of the Australian maritime industry, leading to safer, more efficient, maritime operations, with less environmental impact. However, to achieve that vision, significant effort from Government, the maritime industry, and other stakeholders must be invested to put in place the regulatory frameworks, qualifications frameworks, skills base, and port facilities, that are required.
Transitioning from 2021 to the vision for 2040 will require the advancements contained within the diagram below.
These advancements are within Australia’s reach, if a proactive, coordinated effort, led by Government and incorporating industry and the community is enacted.
If this effort is not put in now, for example because of distrust for new technology, fear about the impact on jobs, an inability to depart from ‘the way it has always been done’, or simply disinterest from the Australian Government, other countries, particularly those with more developed technological capability, will seize the advantage, and monopolise the opportunities on the table. Leveraging Australia’s talented technologists and innovators, maintaining a strong focus on building sovereign capability through multidisciplinary activities, and a Government-led, multi-domain effort to revamp Australia’s regulatory approach to emerging technology, will position the Australian maritime industry to take full advantage of the spectrum of safety, environment, efficiency, and economic benefits of autonomous systems technology. 
“In other words, to fully realise the potential of autonomous shipping, the development technologies must be deemed valuable by the wider marine industry as well as the society as a whole.” (Advanced Autonomous Waterborne Applications – AAWA initiative)
What is TAS doing that supports the vision identified in this paper?
In addition to facilitating the development of game changing trusted autonomous systems technology, Trusted Autonomous Systems (TAS) also has two common-good activities: A1 Ethics and Law of Trusted Autonomous Systems and A2 Assurance of Autonomy. These activities, funded by Queensland Government, provide support and resources to TAS participants and broader commercial and Defence stakeholders.
Under the A2 Assurance of Autonomy Activity, the National Accreditation Support Facility Pathfinder Project (NASF-P) is delivering the following:
- Development of an Australian Code of Practice for the Design, Construction, Survey and Operation of Autonomous and Remotely Operated Vessels
- Development of an enabling COLREGs compliance framework
- Development of a Body of Knowledge on Assurance and Accreditation of Autonomous Systems
The NASF-P has already:
- Hosted the inaugural Regulator Roundtable and Showcase Event in October 21
- Coordinated and funded delivery of world first Autonomous Marine Systems Fundamentals for Marine Surveyors course
Under the A2 Assurance of Autonomy Activity, the Enabling Agile Assurance of Drones in Queensland project is delivering smart digital regulatory tools to enhance efficiency and communication amongst operators and regulators, facilitating innovation and driving growth in industry.
 Horne, R. (2021). Autonomous and remotely operated vessels 2021 to 2040. MIAL Future Leaders White Paper. Predictions for the Australian Maritime Industry 2040. Maritime Industry Australia Limited. pp.12-27
Rachel Horne, Assurance of Autonomy Activity Lead, Trusted Autonomous Systems (TAS)
Autonomous systems offer the ability to increase safety and efficiency while lowering environmental and economic costs. In the last five years, there has been a rapid acceleration in the capacity and availability of autonomous and remotely operated vessels. Autonomous systems need to be trusted by government, regulators, operators and the broader community if this rapid acceleration is to continue and to ensure this technology can integrate into commercial and defence operations. An integral part of gaining trust is having consistent assurance requirements, and a clear, tailored regulatory framework.
The Queensland Government has identified the safety, environmental, and economic opportunities of addressing the assurance and accreditation challenges for autonomous systems, resulting in funding for the Assurance of Autonomy Activity through Trusted Autonomous Systems (TAS). This Activity: “…aims to unlock Queensland’s, and by extension Australia’s, capacity for translating autonomous system innovation into operational capability, leveraging regulatory and technical expertise and strong stakeholder relationships to support industry and regulators.”
To support the development of a clear, tailored, regulatory framework, TAS is leading the development of an Australian Code of Practice for the Design, Construction, Survey and Operation of Autonomous and Remotely Operated Vessels (Australian Code of Practice), supported by Maaike Vanderkooi of Vanderkooi Consulting. The Australian Code of Practice will represent best practice and is intended to provide certainty for industry by providing clear standards, tailored for the type of vessels and operations common in Australia. The Australian Code of Practice will be voluntary and will be updated periodically. It is hoped that the Australian Maritime Safety Authority (AMSA), which has been closely consulted throughout this project, will recognise the Code and in the future consider incorporating it into their regulatory framework.
There are several codes, standards, and guidelines already available internationally for autonomous and remotely operated vessels. The first step in the development of an Australian Code of Practice is understanding the leading existing codes, standards and guidelines, considering them in an Australian context, and then determining whether any of these documents, or specific approaches they take, could be tailored for use in an Australian context. Recognising the importance of this step, TAS engaged Maaike Vanderkooi of Vanderkooi Consulting to prepare a Report (“Analysis of Available Standards and Codes for Autonomous and Remotely Operated Vessels”), as the first stage in the project to develop the Australian Code of Practice.
The Report analyses:
- the UK Code of Practice for Maritime Autonomous Surface Ships;
- the LR Code for Unmanned Marine Systems; and
- DNV GL’s Autonomous and Remotely-operated Ships Class Guideline.
As part of the analysis of these documents, the Report seeks to:
- understand the structure and requirements of each of the codes;
- identify the differences and similarities between the codes; and
- consider the codes in the Australian regulatory context.
The below is extracted from pages 81-82 of the Report
“The report finds that:
- an Australian Code of Practice for autonomous and remotely operated vessels should align with the regulatory framework that already exists for conventional domestic vessels;
- the three available codes focus largely on vessels which comply with international conventions or Class Rules; and
- this is different to the context for an Australian Code of Practice, which will be tailored towards commercial vessels operating only in Australian waters.
For this reason, none of the available codes and standards considered in this report provide a template that could be tailored for use in Australia with only minor modifications.
However, each of the three available codes will significantly influence the content of the Australian code. This report uses the analysis of the three available codes and standards to identify the standards or requirements that should apply to autonomous vessels, beyond the requirements of conventional vessel standards. This will include tailored requirements for:
- situational awareness;
- control systems;
- software integrity and testing; and
- safe states.
This report also finds that the operational requirements that apply to conventional vessels in Australia should apply to autonomous and remotely operated vessels, with some differences:
- the safety management system requirements need to be tailored to autonomous and remote vessel operations;
- the minimum crew and crew competency requirements will need to be modified; and
- there will be additional requirements for contingency planning and control hierarchies, which should be informed by the content of the three available codes and standards.
In line with the available codes and standards, a risk analysis approach, which focuses on the impact of potential failures, should apply to the development and testing of novel systems on the vessel, including the systems for situational awareness and control and all systems which do not meet the requirements of the conventional vessel standards.
Finally, this report notes that the baseline requirement of each of the available codes and standards is for an autonomous or remotely operated vessel to be ‘as safe as’ a conventional vessel. Given that the scope of the Australian code will include very small, low risk autonomous marine equipment, whether or not this baseline approach is appropriate for all vessels subject to the Australian code will need to be considered as part of the consultation process on the development of the Australian code.”
Since this report has been completed, the project has identified the underpinning principles for the draft Australian Code of Practice, using stakeholder consultation to refine and ensure they are fit for purpose (complete) and drafted the Australian Code of Practice (complete).
|1. Undertake consultation on the draft Australian Code of Practice with the Australian Maritime Safety Authority (underway);
2. Undertake public consultation on the draft Australian Code of Practice to ensure it is fit for purpose and will be as useful as possible for Australian industry (expected mid-November 2021); and
3. Finalise and release edition one of the Australian Code of Practice with accompanying guidance material (expected early 2022).
TAS intends to release edition one of the Australian Code of Practice with accompanying guidance material via our website and it will likely also appear directly on AMSA’s website. Feedback on the Australian Code of Practice can be provided directly to either TAS or AMSA to inform continued iterative development.
If you would like to contact us in relation to the TAS Code of Practice project, to offer feedback, suggestions, or request more information, please email us at firstname.lastname@example.org.
 DNV GL, 2018, DNV class guidelines: Autonomous and remotely operated ships, DNVGL-CG-0264
 Information on these principles and the consultation undertaken available here: Vanderkooi, M. and Horne, R., 2021, Outcomes of a successful webinar on TAS’s project to develop an Australian Code of Practice in 2021
On 6 and 7 October, Trusted Autonomous Systems (TAS) will virtually host the Trusted Autonomous Systems Regulator Roundtable and Showcase Event with senior executives from commercial and defence autonomous systems regulatory bodies. Participant include representatives from the Australian Maritime Safety Authority (AMSA), Civil Aviation Safety Authority (CASA), Defence Aviation Safety Authority (DASA), Department of Infrastructure, Transport, Regional Development and Communications (Infrastructure), Maritime Safety Queensland (MSQ), National Transport Commission (NTC), the Office of the Defence Seaworthiness Regulator (ODSwR) and Queensland Government. The roundtable will also be briefed by representatives of Australian Institute of Marine Sciences (AIMS) and QinetiQ on maritime and aerial test facilities for autonomous systems.
TAS are hosting this event as part of its broader efforts to support the growing Australian autonomous systems ecosystem by identifying and addressing key hurdles in the assurance and accreditation frameworks for autonomous systems and connecting stakeholders to broaden understanding of issues and effect meaningful change.
Aims of the Trusted Autonomous Systems Regulator Roundtable and Showcase Event are:
- Facilitating increased awareness amongst commercial and defence regulators of the growing autonomous systems industry in Australia,
- Encouraging shared conversations and understanding of key regulatory issues, opportunities, and approaches, between commercial and defence regulators and key Commonwealth and Queensland agencies,
- Facilitating increased awareness of TAS’ work to support the assurance and accreditation frameworks for autonomous systems and the role of Queensland testing facilities,
- Increasing prioritisation by regulators of autonomous systems related regulatory, policy, and operational requirements and issues, and
- Building momentum for future regulatory pathways for autonomous systems in Australia across all domains.
If you want to know more or want to contribute to the work being undertaken by TAS you can contact the TAS Assurance of Autonomy team on NASFP@tasdcrc.com.au
The synthesis of swarms has evolved over 30 years from simple rule-based ‘artificial life’ simulations to reports of increasingly complex, cognitive, and numerous small autonomous platforms. Despite this progress, technical challenges and fundamental questions remain. We propose to take stock of progress, and seek answers to questions such as: How is agency and risk traded for the individual and the collective? How might swarms use higher-order symbolic and semantic reasoning? How might human operators govern such systems, and ensure behaviours are bounded and ‘controlled’? How might swarms, teams of robots and hybrid swarm-teams allocate and coordinate action, distribute labour, observe, and manifest hierarchies, and self-monitor global properties and goals? What ethical, legal and safety frameworks and governance should ‘thinking’ swarms comply with? There is an opportunity to progress answers to these questions through presentations at invited workshops and subsequent production of a peer reviewed edited book.
A set of virtual and/or physical workshops Q1 2022 led by Trusted Autonomous Systems and UNSW Canberra. Each workshop session will run for 60 minutes, 3 x 10 minutes presentation + 30 min panel Q&A. Presenters and workshop themes to be determined through an EOI process with a workshop program announced by Fri 26 Nov 2021. Workshop presentations will be invited to submit a chapter of a peer reviewed edited book.
The aim of the workshops is to bring together a broad range of experts (including technical and engineering; cognitive science and artificial intelligence; ethical, legal, and safety) to consider technical challenges and fundamental questions about ‘thinking swarms’.
The purpose of the workshops is to co-contribute content to shape an Academic book ‘Thinking Swarms: Agency, teaming, emergence and governance’
- Build a ‘Thinking Swarms’ Community of Practice
- Develop a conceptual framework, themes, and topics to frame the Thinking Swarms edited book
- Invite a set of authors to write chapters for the Thinking Swarms edited book
If you are interested in presenting at the workshops, please email presentation title, 150 word abstract and 100-word bio to the organisers (now extended) to Friday 5 November:
|Kate Devitt, Kate.Devitt@tasdcrc.com.au
Jason Scholz, Jason.Scholz@tasdcrc.com.au
|Simon Ng, Simon.Ng@tasdcrc.com.au
Hussein Abbass, H.Abbass@unsw.edu.au
By Robert Dickie1and Rachel Horne2
1Group Leader, Systems Safety & Assurance, Frazer Nash Consultancy Ltd
2Assurance of Autonomy Activity Lead, Trusted Autonomous Systems (TAS)
Autonomous vessels of various sizes, forms and speeds are already at sea, on the surface and beneath it. The International Regulations for Preventing Collisions at Sea (COLREGs), published by the International Maritime Organization (IMO) in 1960 and updated in 1972, govern the ‘rules of the road’ at sea. COLREGs describe the features that vessels must have to facilitate being seen and identified, define means of communication between vessels for the purposes of signaling intent, and most importantly they describe the navigational behaviors expected of vessels in proximity to one another, for the purposes of avoiding collision. It’s clear from the terms and phrases used in COLREGs that the authors didn’t conceive of navigational or operational decisions being made by computers, and compliance for autonomous vessels is difficult and not well-understood.
Autonomous systems technology does not replicate humans, it emulates some of their skills using a different set of ‘senses’ and decision-making processes and brings new capabilities to operations. This means that humans and autonomous technologies have different weaknesses, strengths, risks and mitigations.
The autonomous maritime industry has been wrestling with the challenge of ‘compliance’ with COLREGs for years, in terms of both understanding how it applies, and how to demonstrate compliance. The challenge for the designer or operator of an autonomous vessel is that the regulations are phrased from the underlying assumption that a human is operating the vessel. Where an autonomous control system is performing some or all of the functions a human previously would have been, it can be difficult to work out what constitutes ‘compliance’, in a practical sense, and in a way that the regulator, the Australian Maritime Safety Authority (AMSA), would accept. This difficulty can lead to additional costs, delay, and operations which are subject to more limitations than may be reasonable based on the actual risks presented.
Developing one-off COLREGs compliance cases for a single autonomous vessel is onerous for the designer or operator, and also causes AMSA difficulty in terms of the resources required to assess the compliance case, and ensure consistency in regulatory decision making. There are significant efficiencies to be gained for designers, operators, and regulators of autonomous vessels in the development of a repeatable compliance framework designed to reduce these burdens.
New TAS project to address COLREGs challenge
The TAS Assurance of Autonomy team have commenced a new project aimed at addressing the COLREGs challenge by developing an enabling framework which supports a practical and appropriate level of compliance for autonomous vessels. The project, known as the ‘TAS COLREGs project’, will consider the specific risks posed by the full spectrum of autonomous vessels, thus allowing flexibility in the operator’s approach to compliance. In order to provide maximum usefulness and future proofing, the approach will be to offer a range of risk mitigation options which address the ‘spirit’ of COLREGs, rather than defining a set of performance requirements or specification with which all autonomous vessels much comply.
The guiding principles for the TAS COLREGs project (see Figure 1) are:
- The compliance methodology will be repeatable and scalable across a broad range of autonomous vessels.
- The underpinning philosophy of the compliance methodology will be logical, reasoned and justified by argument.
- The compliance methodology will be enabling rather than constraining whilst upholding the purpose and spirit of COLREGs.
- The process followed to use the compliance methodology will be simple to follow and supported by guidance.
- The project aims to develop a methodology which can be agreed to by the regulator.
Outputs of the TAS COLREGs project includes an operational tool and user guidance.
Stakeholder engagement will occur over the coming weeks to ensure the tool is fit for purpose and considered acceptable and appropriate by AMSA. This tool will be trialed with a variety of autonomous vessels in 2021, and then refined as necessary, before being released.
The TAS COLREGs project is being delivered by Frazer-Nash Consultancy and led by Robert Dickie, with the support of Marceline Overduin and Andrejs Jaudzems. Robert has extensive experience in the assurance of maritime autonomous systems, having been a member of the UK Maritime Autonomous Systems Regulatory Working Group (MASRWG) and having developed the initial draft of the Lloyd’s Register Code for Un-crewed Marine Systems.
The output of the TAS COLREGs project will be available on the TAS website as a useful resource for designers and operators in the form of a repeatable COLREGs compliance framework, supported by a tool and user companion guide.
The COLREGs resource will also be included in the Body of Knowledge Toolbox being developed by the TAS Assurance of Autonomy team under the National Accreditation Support Facility Pathfinder Project (NASF-P), which will assist designers, operators, regulators, and other stakeholders in the Australian autonomous systems ecosystem to navigate the assurance and accreditation process for autonomous systems efficiently and successfully.
If you would like to contact us in relation to the TAS COLREGs project, to offer feedback, suggestions, or request more information, please email us at NASFP@tasdcrc.com.au.