C-SHARE: Joint ATM Cognition through Shared Representations.
Delft University of Technology (TUD)
National Aerospace Laboratory (NLR)
Thales Nederland (TNL)
It is to be expected that the task of an air traffic controller will change with the introduction of 4D (space and time) trajectories for aircraft. Here, a higher level of automation will play a central role. From various other complex socio-technical domains it is known that automation can often introduce new problems, which can sometimes be more difficult to mitigate than the problems intended to be solved by automation in the first place. C-SHARE aims to provide a framework for a complete functional airspace representation (that is, model of the work domain) that underlies both the design of human-machine interfaces as well as ground and air-automation tools to resolve these problems by design.
Introduction and Problem Statement
Higher levels of automation play a central role in the recent developments within the field of Air Traffic Management (ATM). Increasing the level of automation in ATM, however, can either bring success or failure, depending on how well it supports the human operator in familiar, unfamiliar and perhaps even unanticipated situations. Here, important lessons can be learnt from flight deck automation. The high level of flight deck automation has certainly helped to increase the level of safety and overall system performance, however, the approach has also led to an increased complexity of the operations and significant problems regarding crew situation awareness, workload, over-reliance on automation, skill degradation, and vigilance problems. In hindsight, taking a different representation of the cognitive work to be performed on the flight deck as a starting point to automate the processes involved, might have resolved these problems by design. Clearly then, when introducing higher levels of automation in ATM, one should first of all investigate what representations of the work domain matches the cognition performed by humans and/or automation best.
One of the tasks in which the human is foreseen to remain to play an essential role is in 4D trajectory management that entails planning, revising, and monitoring 4D Business Trajectories (BT). In the planning phase, the Shared Business Trajectories (SBTs) should be structured within the airspace such that an optimum overall system performance will be obtained in terms of safety, efficiency and productivity. During execution of the Reference Business Trajectories (RBTs), unforeseen separation provisions, sequencing, new airspace and business needs, weather and changing arrival constraints may result in the inability to closely adhere to the planned RBTs. Without proper tools to re-plan and re-negotiate the RBTs, the required adaptations might result in an inefficient and unsafe system. Ideally, a truly cooperative process would be desired, in which automation enables the human operators to function to their full potential (as creative problem-solvers) by keeping them continuously involved in the decision-making process. To enhance operator situation awareness, to gain trust in and reliance on automation, a joint cognitive system would be desired in which the automated systems and the human operators share the same representation of the work environment.
Project objectives and expected results
In this project we aim to develop a novel representation of the tactical and strategic air traffic management tasks - a common shared representation (model of the world) that underlies both the actions of the automation and the cognition of the human actors. This so-called Functional Airspace Representation (FAR) can act as a visual interface for the human operator and provides the principles and input for higher levels of automation such as long- and short-term planning, revising, and monitoring of 4DT’s and highly-advanced conflict detection and resolution algorithms.
By virtue of being based on this shared representation, the output of the automated tools will be compatible with the human’s representation of the work domain, and can be visualized in a human centric manner. This Functional Airspace View (FAV) should provide the human operator a window to both the goals and constraints of the functional airspace as well as the goals and constraints of the automation support. This allows for a shift back and forth across the various levels of automation, from fully manual control to fully automatic control, and also supports transition between SESAR’s unmanaged and managed airspace. Such a shared mental model between humans and automation is crucial for the user’s situation awareness, and trust in and acceptance of the automation.
The main outcome of the project will be a complete representation of the tactical and strategic manipulation of four-dimensional trajectories, a framework that can act as the basis for designing both the human-machine interface and automated tools, in the air and on the ground, from one and the same perspective. Integrated in this representation is an automated tool to support the human with a limited number of meaningful conflict resolution options – depending on the desired level of automation. The human-machine interface and automation will be verified through human-in-the-loop simulations.
Approach / methodology
A Work Domain Analysis (WDA) will be performed on the existing airspace structures and the constraints that govern 4D trajectory management. The goal-relevant constraints and their couplings (on different levels of abstraction) on long, medium and short term horizons will be analysed. Computer algorithms will be used to identify and model the deterministic and probabilistic elements of the airspace complexities.
A human-machine interface (HMI) of a joint-cognitive system containing a shared representation that can guide both the design of decision-support human-machine interfaces as well as ground- and air-automation tools will be developed. The HMI will provide a radically different way to represent the (manipulation of) four-dimensional trajectories. A shared representation for automation and interface yields significant advantages, as it allows us to shift back and forth across the various levels of automation. i.e., from fully manual control to fully automatic control, and also from transitioning between SESAR’s unmanaged to managed airspace. The design of the joint cognitive system will be guided by following the phases of a Cognitive Work Analysis (CWA).
Extensive human-in–the loop experiments will be conducted to evaluate the HMI in terms of automation support for various scenario’s (e.g., transitions from unmanaged to managed airspace, high/low traffic densities, long-term and short-term planning, last-minute deconflictions (traffic, weather), etc.) as well as human factors (e.g. situation awareness, workload, decision-making, trust, acceptance, etc.). The evaluations will follow the INTUITION user-centred process that is used within the Thales Human Factors and Cognition Lab and will be used in a bootstrapping approach within the CWA approach.
Finally, project dissemination activities will be including scientific publications, press releases, and workshops with domain experts and connection with stakeholders (including industry and ATCos).