Driving Change in Aviation: The Role of Data Science and Human-Machine Collaboration

Paula Lopez

2023-09-05 09:36:54
Reading Time: 2 minutes

It has been a year since the launch of the SafeTeam project to our esteemed readers at datascience.aero. Innaxis, a pioneer in data science in aviation with notable projects like ComplexWorld, Prospero, and SecureDataCloud, has consistently advanced its research in this field. Building upon this accumulated knowledge, DataBeacon has taken this research further, developing digital assistants and data-driven applications that augment human operations and enhance safety and performance in the air transport system.

Despite the demonstrated potential and challenges of these technologies, there are still barriers to their implementation in aviation, particularly in air traffic management (ATM). These barriers are inherent to the complex nature of the ATM sector and market and are resistant to change. However, collaboration between key stakeholders—such as human factors experts, regulatory agencies, technology developers, and end users—can help mitigate some of these challenges. This collaborative effort is precisely what we have undertaken in SafeTeam.

Over the past decade, automation technologies, including machine learning (ML), have witnessed rapid advancements, finding applications across various sectors like manufacturing and autonomous driving. However, in aviation, the adoption of new technologies is considerably slower due to the imperative of maintaining current safety levels. And, with the need to continue promising said safety levels despite current levels of air traffic, limitations in scalability and resources, workload peaks and growing demand, technological support will be even more important in the near future. Achieving the safe implementation of digital assistants requires extensive research on human factors and revised regulatory frameworks. While the need for such research is common to all digital assistance developments, each application in aviation demands dedicated investigation. Consequently, SafeTeam has chosen a use case-driven approach, concentrating on three specific use cases for which the technology was already mature, allowing for testing, adaptation to user needs, and iterative validation.

During the past year, the team has focused on the definition phase of the project. Two use cases have been meticulously defined from technological, human factors, and operational perspectives: an en-route digital assistant for air traffic controllers led by DataBeacon and its Victor5 Digital assistant and a UA (Unstable Approach) digital assistant for pilots in the cockpit led by TUM and its cockpit simulator. By closely examining each use case, the project has addressed key questions regarding human-machine teaming to comprehensively characterise the user, the technology, and the interaction between both entities. This thorough understanding enables us to maximise the potential of this human-machine collaboration and unlock the untapped opportunities inherent in such partnerships.

Ri.Se has made significant contributions to the project by developing a comprehensive framework aimed at guiding technology developers in incorporating human factors considerations throughout the various design phases. This formal framework encompasses idea generation, system modeling, allocation model, and implementation. By adhering to these guidelines, developers can effectively address critical questions during system modeling and design. These questions encompass not only the users but also the agents and organisations involved in the proposed changes, their respective tasks, potential impacts on stakeholders, and existing regulatory frameworks, among others. Additionally, this framework assists in defining the human and machine functions necessary to support the allocation model and determine the appropriate level of autonomy. By following this model, developers can identify and evaluate the strengths and weaknesses of both humans and machines, as well as those of the new system resulting from their collaboration, utilising the revised HABA MABA model.

All research outcomes and findings will be made publicly accessible on the project website very soon.

Author: Paula Lopez

© datascience.aero