Periodic Reporting for period 1 - SafeTeam (Safe Transition to Digital Assistants for Aviation)
Período documentado: 2022-07-01 hasta 2023-12-31
The SafeTeam project arises in a changing aviation environment, where new technology meets traditional industry methods focused on a strong and necessary safety culture.
Rooted in Paul Fitts' pioneering work of 1951, which delineated the roles of humans and machines in task distribution, SafeTeam analyzes the role of automation in reshaping aviation operations and investigates on its safe integration. Over the decades, from the advent of AI research in the 1950s to the recent surge in algorithmic training fueled by big data and cloud computing, aviation has witnessed a gradual integration of automation.
Today, on the cusp of a paradigm shift, SafeTeam seeks to harness the transformative potential of AI and digitalization to enhance safety and efficiency in aviation. Leveraging concrete case studies within aircraft crew and air traffic control operations, SafeTeam aims to propel the industry towards higher levels of automation while prioritizing safety and human-centric design. These case studies, meticulously selected based on technological maturity and industry relevance, serve as testbeds for evaluating the efficacy of intelligent assistance systems.
Central to SafeTeam's mission is a deep understanding of human factors and safety considerations in the context of automation. By assessing AI performance metrics such as explainability and accuracy, alongside considerations of workload and situational awareness, SafeTeam aims to optimize human-machine interaction and ensure seamless integration of automation into operational workflows.
The objectives of SafeTeam are manifold, encompassing technological advancement, regulatory compliance, and industry collaboration. By championing a human-centric approach to automation, SafeTeam aims to facilitate the seamless integration of AI technologies into aviation operations, underpinned by robust methodologies for performance assessment and monitoring. Furthermore, SafeTeam seeks to support the development of digital assistants for aviation operations while ensuring alignment with regulatory and certification requirements.
In essence, SafeTeam represents a concerted effort to navigate the complexities of automation in aviation, guided by a commitment to safety, collaboration, and innovation. By forging interdisciplinary partnerships and embracing a holistic approach to research and development, SafeTeam endeavors to shape the future of aviation in a manner that enhances both human performance and industry resilience.
Rooted in Paul Fitts' pioneering work of 1951, which delineated the roles of humans and machines in task distribution, SafeTeam analyzes the role of automation in reshaping aviation operations and investigates on its safe integration. Over the decades, from the advent of AI research in the 1950s to the recent surge in algorithmic training fueled by big data and cloud computing, aviation has witnessed a gradual integration of automation.
Today, on the cusp of a paradigm shift, SafeTeam seeks to harness the transformative potential of AI and digitalization to enhance safety and efficiency in aviation. Leveraging concrete case studies within aircraft crew and air traffic control operations, SafeTeam aims to propel the industry towards higher levels of automation while prioritizing safety and human-centric design. These case studies, meticulously selected based on technological maturity and industry relevance, serve as testbeds for evaluating the efficacy of intelligent assistance systems.
Central to SafeTeam's mission is a deep understanding of human factors and safety considerations in the context of automation. By assessing AI performance metrics such as explainability and accuracy, alongside considerations of workload and situational awareness, SafeTeam aims to optimize human-machine interaction and ensure seamless integration of automation into operational workflows.
The objectives of SafeTeam are manifold, encompassing technological advancement, regulatory compliance, and industry collaboration. By championing a human-centric approach to automation, SafeTeam aims to facilitate the seamless integration of AI technologies into aviation operations, underpinned by robust methodologies for performance assessment and monitoring. Furthermore, SafeTeam seeks to support the development of digital assistants for aviation operations while ensuring alignment with regulatory and certification requirements.
In essence, SafeTeam represents a concerted effort to navigate the complexities of automation in aviation, guided by a commitment to safety, collaboration, and innovation. By forging interdisciplinary partnerships and embracing a holistic approach to research and development, SafeTeam endeavors to shape the future of aviation in a manner that enhances both human performance and industry resilience.
In the first half of the project, SafeTeam completed the Definition phase, establishing the foundation for the Implementation and Validation phases in the latter part of the project. Specifically:
- WP2 outlined generic design principles facilitating coordination between human and digital assistants. The SafeTeam framework offers a methodological approach tailored for researchers and practitioners with limited or no experience in Human Factors and Human-automation interaction. It aims to identify the nature and format of information required by digital assistants to optimize human-machine cooperation. Ultimately, the goal is to design an intelligible digital assistant that maximizes teamwork with human operators. To achieve this, SafeTeam investigates designing explanations to support human-machine cooperation, aligning them with cognitive aspects. This research draws on recent insights from ergonomics, social robotics, and psychology concerning the mechanisms underlying cooperative action control.
- WP3 concentrated on defining two digital assistants for aviation operations: one for en route air traffic control and another for unstabilized approach (UA) alarms in the cockpit. Both use cases underwent comprehensive definition from operational, organizational, and technological perspectives in collaboration with operational experts (pilots and air traffic controllers). This collaboration aimed to design digital assistants safely integrable into current operations while streamlining them. The en-route ATC digital assistant enhances airspace planning and supervision efficiency, while the UA Digital Assistant adds an additional safety layer to landings.
- WP2 outlined generic design principles facilitating coordination between human and digital assistants. The SafeTeam framework offers a methodological approach tailored for researchers and practitioners with limited or no experience in Human Factors and Human-automation interaction. It aims to identify the nature and format of information required by digital assistants to optimize human-machine cooperation. Ultimately, the goal is to design an intelligible digital assistant that maximizes teamwork with human operators. To achieve this, SafeTeam investigates designing explanations to support human-machine cooperation, aligning them with cognitive aspects. This research draws on recent insights from ergonomics, social robotics, and psychology concerning the mechanisms underlying cooperative action control.
- WP3 concentrated on defining two digital assistants for aviation operations: one for en route air traffic control and another for unstabilized approach (UA) alarms in the cockpit. Both use cases underwent comprehensive definition from operational, organizational, and technological perspectives in collaboration with operational experts (pilots and air traffic controllers). This collaboration aimed to design digital assistants safely integrable into current operations while streamlining them. The en-route ATC digital assistant enhances airspace planning and supervision efficiency, while the UA Digital Assistant adds an additional safety layer to landings.
The primary final result achieved is the SafeTeam framework. This framework establishes guidelines and design principles to facilitate cooperation between humans and automation. It is based on a comprehensive literature review of evidence-based methodologies, which have been synthesized and restructured to enable individuals without expertise to design and/or evaluate their automated systems with human-machine collaboration in mind. The proposed framework is divided into three main phases, with encouragement for iterations within these phases. Phase I emphasizes system understanding, utilizing Hierarchical Task Analysis and interviews. Phase II focuses on safe human-autonomy teaming and introduces a set of design principles supporting continuous consideration of autonomy teaming during system design. The third phase proposes an assessment of the design proposal to mitigate collaboration issues and risks, guiding the evaluation through a set of questions.
The definitions of use cases are regarded as intermediate results that will facilitate their implementation and validation in the second phase of the project. These use cases are built upon existing testing platforms, allowing for rapid prototyping and relevant validations. Specifically, the enroute Air Traffic Control (ATC) use case is developed on Victor5, created by DataBeacon, an innovative AI digital assistant designed to enhance ATC efficiency by providing real-time surveillance data and facilitating interactions among air traffic.
Within Victor5, SafeTeam is implementing a complexity metric capable of real-time measurement of the complexity of a given airspace sector. This metric advances the current state of the art by incorporating potential interactions among aircraft in its assessment of complexity. This complexity directly impacts Air Traffic Control Operator (ATCO) workload and is thus considered a valuable tool for sector planning, demand-capacity balance, and post-analysis. The Unstable Approaches Digital Assistant is built upon a research cockpit simulator available at TUM. This simulator offers flexibility to test various Human-Machine Interface (HMI) designs and scenarios. Several cockpit crews will be invited to participate in simulator scenarios, allowing for assessment of human-machine performance. The results will serve as a valuable source for adjustments and adaptations of the digital assistant and its HMI.
The definitions of use cases are regarded as intermediate results that will facilitate their implementation and validation in the second phase of the project. These use cases are built upon existing testing platforms, allowing for rapid prototyping and relevant validations. Specifically, the enroute Air Traffic Control (ATC) use case is developed on Victor5, created by DataBeacon, an innovative AI digital assistant designed to enhance ATC efficiency by providing real-time surveillance data and facilitating interactions among air traffic.
Within Victor5, SafeTeam is implementing a complexity metric capable of real-time measurement of the complexity of a given airspace sector. This metric advances the current state of the art by incorporating potential interactions among aircraft in its assessment of complexity. This complexity directly impacts Air Traffic Control Operator (ATCO) workload and is thus considered a valuable tool for sector planning, demand-capacity balance, and post-analysis. The Unstable Approaches Digital Assistant is built upon a research cockpit simulator available at TUM. This simulator offers flexibility to test various Human-Machine Interface (HMI) designs and scenarios. Several cockpit crews will be invited to participate in simulator scenarios, allowing for assessment of human-machine performance. The results will serve as a valuable source for adjustments and adaptations of the digital assistant and its HMI.