Jacqueline A. Duley and Raja Parasuraman The Catholic University of America
Over the last several decades the aircraft flightdeck has evolved in its use of automation. The earliest implementations of automation involved providing assistance to the pilots in the form of maintaining aircraft control. For example, the autopilot was developed because of the inherent instability of the aircraft during flight. With the introduction of electronic displays, automation provided assistance to pilots in navigating their aircraft with the addition of the plan view display, which incorporated much of their flight planning information into one integrated display. Fadden ( 1990) termed these two types of aviation automation, control and information automation, respectively.
There have been considerable benefits to aviation from the introduction of these and other forms of flight deck automation ( Billings, 1997). Nevertheless, several incidents and accidents have demonstrated the problems that can arise when poorly-designed automation is implemented (Parasuraman & Riley, 1997). Some forms of automation which Wiener ( 1989) has termed "clumsy" have resulted in the pilot often being off-loaded during periods of low workload and being given additional tasks to perform in periods of high workload. As a result, the anticipated workload reducing benefits of automation have not been realized. Other documented problems of some forms of cockpit automation include mode errors, opaque interfaces, and automation complacency (for reviews, see Parasuraman & Riley, 1997; Sarter, 1996).
In contrast to the flight deck, automation of air traffic control has not yet reached a high level of complexity and authority ( Hopkin, 1995; Wickens, et al., 1997). However, large projected increases in air traffic are driving several efforts to introduce advanced automation tools for controllers in the National Airspace System (NAS). Currently, air traffic controllers working en-route sectors use plan view displays to guide aircraft traversing the sectors en-route to the receiving Terminal Radar Approach Control (TRACON). In an en-route sector, the controller monitors the ongoing flights, responds to various pilot requests, and adjusts to weather conditions or other anomalies by instructing pilots to alter their aircraft airspeeds, flight levels, and often headings in order to maintain safe and efficient flow of traffic throughout the current as well as adjacent sector(s) ( Wickens, et al., 1997). Under this air traffic control philosophy, the controller has complete authority and responsibility for maintaining the minimum allowable separation among aircraft in the assigned sector.
The volume of air traffic is projected to increase by up to 100% over the next decade. As a result, increasing capacity while maintaining or enhancing safety has become a priority for air traffic management. Responses to this increased demand such as constructing new airports, improving aircraft flow management, or expanding existing facilities, can only provide a partial solution because of their limited applicability or high cost. Two other responses have been proposed ( Wickens, et al., 1998). The first response to such demands is to increase the use of automation in the air traffic control environment in order to aid controllers under high traffic load conditions. A recent report by the National Research Council Panel on Human Factors in ATC Automation recently noted the following as areas of potential controller vulnerability: monitoring for and detection of unexpected low-frequency events; expectancy-driven perceptual processing; extrapolation of complex four-dimensional trajectories; and use of working memory to either perform complex cognitive problem solving orto temporarily retain information ( Wickens, et.al, 1997). This proposal