Transitioning basic research to build a dynamic model of pilot trust and workload allocation

For pilots to accept active decision aids during complex flight scenarios, it is essential that the automation work in synergy with the aircrew. To ac complish this, the automation must go well beyond menu and macro selections , where the pilot explicitly tells the automation what to do and when to do it. It must also transcend "mother may I" approaches, where the automation asks for permission to proceed. To breach these traditional barriers, the automation needs a sense of how the pilot will react in a given situation a nd, based on that reaction, how much of the workload could be allocated to the automation at any given time. To advance the applied aspects of the res earch in this area, the authors transitioned results from three major resea rch efforts. We implemented a dynamic model of pilot trust and workload all ocation that correlates with the human data collected during those efforts. This "full degrees of freedom" engineering model is an early attempt to bu ild a continuous adaptive process that divides cockpit workload between hum an and machine, and that minimizes human cognitive overload. This paper pro vides implementation detail, and relates those details to the original rese arch and human factors literature. An example of the model's use based on a n involved flight scenario is presented. (C) 1999 Elsevier Science Ltd. All rights reserved.