The need for a more safe and efficient air transportation system with ever increasing requirements of airspace capacity will demand for adaptively automated aids. This in order to keep the workload of the human operators acting in such a complex work environment acceptable. In that scenario, rich information coupling should be provided for a proper human-machine coordination, while taking care of providing the human operator with the most relevant information. The problem, however, is that the notion of relevance – and therefore the appropriate level of automation – changes over time due to the changing context. Furthermore, adaptively automated aids are rarely present in operational systems due to the lack of clear principles on how to design them (Parasuraman, Sheridan, and Wickens 2000).
Typically, the approach for designing adaptively automated aids has relied on the use of psycho-physiological measures to evaluate the operator’s state, or the use of different kind of performance metrics and behavioral models (Kaber and Endsley 2004). This kind of user-based design contributes to the understanding of the human performance during certain control tasks. However, this understanding is far from being enough for designing adaptively automated aids, and that is demonstrated with the nonexistence of a design framework for such purpose.
The aim of this research project is to design and test an adaptively automated ATC aid able to switch and transition between levels of automation in an objective, reliable, and unambiguous way without compromising safety, performance, and operator situation awareness.
The main focus of this project lies on the rerouting control task, mainly because extensive research on automated aids for this purpose can be found; and because it entails the conflict detection and resolution task, which is now and in the future the core task of ATC. The task and activities that drive the controller’s work has been selected as the limiting scope of this research, since our main focus is the conflict detection and resolution task. The other sub-categories of context will not be taken into consideration.
Reach and understanding of the most recent advances in the fields of Adaptive Automation, Levels of Automation and Context-Aware Application Design.
Search for objective metrics that could be used to best represent the task context in which the controller is working.
Identification of the context elements responsible for the difficulty of the conflict detection and resolution task.
Develop a context-sensing methodological framework that can identify task contexts that need automated support.
Design unambiguous and reliable levels of automation that can assist the controller during the most relevant task contexts.
Implement an experimental simulator that adaptively changes the level of automation based on the identified task context.
Evaluate the human-machine system in terms of safety, performance, and operator’s situation awareness, reliability, and acceptance.
The project intends to contribute with empirical research in the field of adaptive automation by making use of a work domain based approach. An analysis of what needs to be done, instead of what the operator state is after doing it, will be the basis of the approach.
Review on the state of the art in the fields of Adaptive Automation, Levels of Automation and Context-Aware Application Design.
Design of a context-sensing application
A context-aware application should be able to identify the task(s) with which the operator is busy. Therefore, an algorithm able to make such inference will be needed, and, in order to achieve objectivity, it will only be based on metrics extracted from the traffic situation. This phase will be elaborated in three research steps: the development of a methodological framework for sensing context, a prototype application, and a validation framework.
The methodological frameworks developed within the discipline of Cognitive Systems Engineering conforms the basis for the first research step. The Control Task Analysis (second stage of a Cognitive Work Analysis) yields a list of events that need to take place in order to achieve the main controlling goal. Kilgore, St-Cyr, and Jamieson (2009) elaborated such analysis for the ATC rerouting task, in which seven subtasks were identified. Furthermore, for the identification of relevant context parameters, Huang and Gartner (2009) propose, for every subtask, to identify the relevant features by analyzing the action’s goal and the software requirement specifications. A description of the possible implementations for achieving the action’s goal can help identify the relevant context parameters.
The current research has so far identified the Separation Monitor (developed by NATS) and the Solution Space Diagram (developed by Delft University of Technology) as the possible sources of relevant context parameters for the rerouting task. Metrics extracted from these diagrams could be used to infer the number of times and the difficulty with which the subtasks identified by Kilgore, St-Cyr, and Jamieson (2009) are being executed by the operator.
This research phase will conclude with a validation of the methodological framework for sensing context. For this purpose, a prototype will be experimentally tested in order to check for correlations between the identified context elements responsible for the task difficulty and instantaneous self-assessments of workload.
Levels of automation design
The objective of designing unambiguous and reliable levels of automation will be achieved during this phase. Having established the methodological framework for identifying the contextual situations that need automated support, an analysis of the operator information needs will be performed. This will serve as input for the level of automation design process, which in turn will be based on the principles of Ecological Interface Design in order to fulfill our objectives of unambiguity and transparency of the levels of automation.
The proposed design will be experimentally evaluated in order to check for compliance with the design objectives. During this experimental phase, a series of test subjects will perform the control task at fixed levels of automation. The identified contextual situation parameters will then be used to confirm whether or not reductions of the effort invested on the tasks for which the level of automation was designed took place. This experimental data will also be used to evaluate the effects the designed levels of automation have on workload, performance and SA.
Experimental simulator implementation
Having designed the levels of automation and established their triggering mechanism by means of a context-sensing application, human-in-the-loop simulations will be conducted in order to test the adaptively automated aid. The task context will continuously be inferred during these simulations, and the level of automation changed accordingly. The ATC simulation tool used for this purpose will be the one developed by Delft University of Technology under the MUFASA WP-E project, sponsored by EUROCONTROL. The interaction of the operator with the simulator will be limited to the tasks identified in the Cognitive Task Analysis.
Statiscal analysis The benefits and drawbacks of the complete context-aware adaptively automated aid will be statistically analyzed in terms of safety, performance and situation awareness. Performance metrics will be defined based on the achievement of the fundamental goals during the simulation trials.
In order to describe the acceptance and reliability of the adaptively automated aid, a qualitative analysis of the responses obtained from the controllers during simulations runs will be elaborated.
Dey, Anind K., Gregory Abowd, and Daniel Salber. 2001. “A Conceptual Framework and a Toolkit for Supporting the Rapid Prototyping of Context-Aware Applications.” Human-Computer Interaction 16 (2) (December): 97–166.http://www.tandfonline.com/doi/abs/10.1207/S15327051HCI16234\_02.
Huang, H, and G Gartner. 2009. “Using Activity Theory to Identify Relevant Context Parameters.” In Location Based Services and TeleCartography II, 35–45. http://link.springer.com/chapter/10.1007/978-3-540-87393-8\_3.
Inagaki, Toshiyuki. 2003. “Automation and the Cost of Authority.” International Journal of Industrial Ergonomics 31 (3) (March): 169–174. doi:10.1016/S0169-8141(02)00193-2. http://dx.doi.org/10.1016/S0169-8141(02)00193-2.
Kaber, David B., and Mica R. Endsley. 2004. “The Effects of Level of Automation and Adaptive Automation on Human Performance , Situation Awareness and Workload in a Dynamic Control Task.” Theoretical Issues in Ergonomics Science 5 (2): 113–153.
Kilgore, R. M., O. St-Cyr, and G.A. Jamieson. 2009. “From Work Domains to Worker Competencies: A Five-Phase CWA for Air Traffic Control.” In Applications of Cognitive Work Analysis. Taylor & Francis.
Kofod-Petersen, Anders, and Jörg Cassens. 2006. “Using Activity Theory to Model Context Awareness.” Lecture Notes in Computer Science (3946): 1–17.
Lagu, Amit V., Steven J. Landry, and Hyo-sang Yoo. 2013. “Adaptive Function Allocation Stabilization and a Comparison of Trigger Types and Adaptation Strategies.” International Journal of Industrial Ergonomics null (null) (April). doi:10.1016/j.ergon.2013.02.006.http://dx.doi.org/10.1016/j.ergon.2013.02.006.
Parasuraman, Raja, and Thomas B. Sheridan. 2008. “Situation Awareness, Mental Workload, and Trust in Automation: Viable, Empirically Supported Cognitive Engineering Constructs.” Journal of Cognitive Engineering and Decision Making 2 (2): 140–160. doi:10.1518/155534308X284417. http://edm.sagepub.com/content/2/2/140.short.
Parasuraman, Raja, Thomas B. Sheridan, and Christopher D. Wickens. 2000. “A Model for Types and Levels of Human Interaction with Automation.” IEEE Transactions on Systems, Man, and Cybernetics. Part A, Systems and Humans : a Publication of the IEEE Systems, Man, and Cybernetics Society 30 (3) (May): 286–97. http://www.ncbi.nlm.nih.gov/pubmed/11760769.
Woods, David D. 1996. “Decomposing Automation: Apparent Simplicity, Real Complexity.” In Automation and Human Performance: Theory and Applications, edited by Raja Parasuraman and M. Mouloua, 3–17. Erlbaum. http://csel.eng.ohio-state.edu/woods/automation/automation94chapter.pdf.