Enhancing Decision-Making Onboard Intelligent Ships

Faster OODA Loops through Human-Machine Teaming

Setting the Stage – The World in 2040

In this DASA funded challenge, the goal was “to revolutionise decision making in a future operating environment onboard an Intelligent Ship in 2040, considering the human-machine teaming aspects and the human operators throughout the design”.

This presented an opportunity to develop a low TRL prototype for a novel user interface in a “Cobiotic Ops Room.” However, before delving into the solution, it was crucial to understand the future operating environment and the challenges that would arise for end-users onboard Navy vessels in 2040.

To understand what the future might be like, extensive research was conducted on Future Naval Surface Vessel Concepts, focusing on the Dreadnaught 2050 and the Thales Tx Ship. A futurist was also interviewed to generate hypotheses about the political, economic, sociological, technological, legal, and environmental trends that would shape the year 2040. Additionally, a thorough scan of Google patents for wearable and social technologies provided further insights into the technological landscape.

Notably, the year 2040 would see the rise of the Centennials, individuals born between 2000 and 2020. This generation would have unique characteristics, being smart, mature, and entrepreneurial, but also exhibiting a short attention span and a reluctance to memorise vast amounts of data. Their proficiency in switching between diverse topics and embracing technology would be a defining trait.

Understanding End User Challenges – Observing an Ops Room

To address the challenge effectively, it was essential to gain a deep understanding of the decision-making processes onboard a Naval Vessel. A day was spent observing training scenarios at HMS Collingwood, documenting the roles, processes, and information flow within the Ops Room.

We wanted to capture the various OODA loops – Observe, Orient, Decide and Act – the four step approach to decision-making that focuses on filtering available information, putting it in context and making the most appropriate decision, while also understanding that changes can be made as more data becomes available.

During the observation, it became evident that the Principal Warfare Officer (PWO) held a critical decision-making role, but also presented a potential bottleneck. Other team members gathered data from various systems, which the PWO used to make decisions. However, the PWO often faced overload when handling many contacts simultaneously. To cope with this high workload, the PWO adopted a “zoom-in” strategy to focus on priority contacts and a “zoom-out” strategy to regain situational awareness when the workload reduced. Unfortunately, this approach could lead to missing indicators of other significant events building whilst attention was focussed elsewhere.

Empathising with the PWO

To address the challenges identified in the Ops Room, the focus was shifted towards understanding the PWO and their decision-making processes in-depth. The Critical Decision Method (CDM), a retrospective interview strategy, was employed to elicit expert knowledge from five PWOs regarding decision-making during safety-critical non-routine threat evaluation and weapons allocation events. This method, combined with self-disclosure, open-ended questioning, and active listening, created a space of trust and openness during the interviews.

Through synthesising these interviews and observations, crucial insights into the PWO’s workload and decision-making criteria were gained. The PWO’s workload was influenced by several factors, including the number of contacts, time constraints (affected by range and speed), situation rate of change (driven by type and speed of threats), number of potential courses of action, complexity of options, certainty (influenced by the environment and sensors), and severity of the threat. Decisions were prioritised on the basis of time constraints and establishing hostile intent through a comprehensive assessment procedure.

Future-casting solutions – Cobiotic Ops Room Vision

With a deep understanding of the PWO’s challenges and decision-making processes, it was time to envision potential solutions. A future casting workshop was arranged where a team of four technical experts back-casted the necessary technology, tools, and processes to address the “data deluge challenge” on an intelligent ship. The output of this workshop was the “Cobiotic Ops Room Vision,” where the PWO would interact with autonomous systems through novel user interfaces.

A hypothesis was formed, suggesting that “Cobiotic Algorithms” could significantly reduce the PWO’s workload. These algorithms would classify contacts based on specific parameters, such as clustering by range, bearing, and speed. This approach would allow the PWO to execute one course of action per cluster instead of handling each potential target individually, thus improving decision-making speed and efficiency.

Engaging End Users – Games to Elicit Reactions

To ensure that the proposed concepts resonated with future end-users, a crucial step was to engage with Centennials, who would form the largest age cohort in 2040. To gather valuable insights and feedback, a game called “Creepy Vs Awesome” was devised. The game involved describing various novel user interface concepts, such as radio-frequency implants, chatbots, avatars, prosthetic devices, and direct brain implants, and recording the participants’ reactions to each concept.

This game was played with the Centennials, as well as with several friends and colleagues, to gather a diverse range of perspectives. From these interactions, key insights were identified:

  1. Agency and privacy were paramount concerns for most individuals, emphasising the need for technology connections to be voluntarily switchable.
  2. The loss aversion bias was observed, meaning people were hesitant to replace their natural limbs with prosthetics unless a limb was lost by accident.
  3. Personalisation of technology was deemed crucial, with most participants expressing an affinity for emoji’s and avatars.
  4. Data security was important, but mainly when information could be used for nefarious purposes like criminal activity.

These insights served as essential criteria for selecting the most suitable concepts to incorporate into the Cobiotic Ops Room design.

Communicating Concepts – Story-Telling with Science Fiction

Recognising the power of storytelling, the insights gained from user engagement were used to create a compelling science fiction narrative and a graphic artist illustrated the story. This story aimed to effectively communicate the vision of the Cobiotic Ops Room onboard an intelligent ship in 2040. To ensure accuracy and authenticity, a Navy Subject Matter Expert (SME) reviewed the terminology. 

The science fiction narrative was then shared with the project team, colleagues, and customers, who responded with enthusiasm and positive feedback. The story served as an engaging way to convey the concepts and foster understanding and excitement about the potential future technology.

The First Iteration – Cobiotic Algorithms and Human-AI Interaction

The next crucial step in the project involved the development of the first iteration of Cobiotic Ops Room solutions. A session was conducted with a data scientist and a software engineer, using the abstraction laddering technique to guide their discussions. The outcome of this collaborative effort was a set of algorithm design principles that ensured the Cobiotic Algorithms supported the concept of human-machine teaming.

Some of the key design principles included adhering to Miller’s Law, which limits the number of clusters to 7+/-2 before overload, using deductive reasoning for decision-making, and designing for flexibility, allowing clustering appropriate to the operational tempo and operator capacity.

Moreover, significant attention was given to the interaction between humans and AI systems. Various ways of displaying contacts were sketched, taking into account gestalt principles and the PWO’s decision-making requirements. Four different display variants were explored:

  1. Radar view with the ship at the centre.
  2. Radar view with the ship at the bottom left.
  3. Threat view, where threats were clustered using Cobiotic Algorithms.
  4. Threat train view, incorporating a time element for enhanced situational awareness.

A low-TRL prototype was built in Excel, allowing end-users to explore the four abstractions of the threat evaluation and weapons allocation scenario and determine the most effective way of presenting contacts.

Final Outcome – Cobiotic Algorithms for Human Autonomy Teaming

Significant progress was achieved in implementing the Cobiotic Algorithms for decision-making. A Navy SME reviewed the visualisations of the scenarios, providing highly positive feedback, particularly on the clustered threat view.

The Cobiotic Algorithms have shown promise and applicability, leading to their adaptation for the DSTL-funded Human Autonomy Teaming for Adaptive Systems (HATAS) project. HATAS, a Dstl funded flagship project, brings together industry stakeholders to assess military concepts, further supporting the Cobiotic Ops Room vision.

Enhancing decision-making onboard Intelligent Ships through faster OODA loops using effective human-machine teaming with Cobiotic Algorithms could transform naval operations in the future.

Leave a Reply

Your email address will not be published. Required fields are marked *