Understanding Flight Deck Task Management: A Prelude to Human-Centered Design

Abstract

Several related concepts, including task management (TM), agenda management, and workload management, have recently been developed to describe a major function performed by pilots on commercial flight decks. This function is related to the fact that real-time operation of complex systems often involves multiple tasks which must be performed concurrently or in rapid succession, and therefore, must be managed. It is not that this function is new. It is more that we are better able, with recent advances in cognitive psychology and associated tools and methods, to identify and quantify it. While TM appears to compose an increasingly large part of the flight crew's role on the flight deck, our understanding of it is in its infancy, and it has been largely overlooked in flight deck function, task, and information requirements analyses. Thus it is not systematically considered in the design process. This symposium explicitly addresses task management as a high level flight deck function. The basic premise is that we must thoroughly understand what TM is, how pilots currently perform it on the flight deck, and what problems occur in pilot-performed TM, before we can address TM issues with a human-centered design approach.

This symposium consists of four papers which seek to describe TM processes and document TM practices and problems in real flight deck environments. The Rogers paper provides normative and operational descriptions of task management that provide useful insights into the set of processes that compose TM and the way that pilots view TM in a real-time setting. His distinction of strategic and tactical TM, based on the time-constraints associated with a particular flight situation, may prove useful in determining the types of design aids that are appropriate under different flight conditions. Pre-planning and contingency planning aids may benefit strategic TM while allocation, prioritization, and memory aids may benefit tactical TM.

The Schutte and Trujillo paper and the Latorella paper provide data from full-mission simulation studies that quantify TM activities and problems in non-normal and normal flight, respectively. The Schutte and Trujillo paper suggests that subjects spend considerable time performing TM during non-normal situations, and those subjects that perform best in fault and mission management tasks seem to use a common rule of thumb (aviate, navigate, communicate, manage systems) to determine the order in which initial monitoring and situation assessment tasks occur, and a situation/event-dependent strategy to determine the order that particular discrete tasks should be performed. Application of inappropriate strategies in ordering tasks appears to contribute to operational errors.

The Latorella paper investigates an aspect of TM, interruption management, by quantifying deleterious effects of interruptions on procedural tasks during descent. She found over 50% more errors (omissions, misorderings, or redundantly performed activities) occur when procedures are interrupted than when not interrupted. She also found that subjects spent a greater proportion of time engaging in unnecessary flight path management during interrupted procedure conditions than uninterrupted conditions. Finally, she reported that rather than slowing procedure performance, interruptions tended to hasten performance on remaining procedural tasks. This could be related to the observation reported in the Rogers paper that pilots seem to use a TM practice of hurrying the pace of task performance during tactical TM, when there are pressing time constraints.

The Funk and McCoy paper describes an on-going effort to: (1) document agenda management errors in aviation accidents and incidents; (2) develop a formal normative model of agenda management; and (3) facilitate agenda management. Funk and McCoy take a broader perspective of task management, re-naming it “agenda management” to include activities performed by humans (i.e., tasks) and functions performed by any goal-directed “actor,” human or automated. Thus agenda management includes management of goals, functions, actor assignments and resource allocations. This is the most formal, complete analysis of task or agenda management activities to date. The level of detail provided here is necessary before human-centered TM aids can be designed.

In combination, these studies begin to provide a basis for understanding the heretofore neglected flight deck function of task management. This understanding, at both a conceptual and an operational level, is required before TM can be systematically addressed within the flight deck design process.