Development and Evaluation of a Cockpit Decision-Aid for Emergency Trajectory Generation

Abstract

The application of intelligent cockpit systems is examined to aid air transport pilots at the task of planning and then following a safe four-dimensional trajectory to the runway threshold during emergencies. The design of a proof-of-concept system is described, including the use of embedded fast-time simulation to predict the trajectory de ned by a series of discrete actions, the models of aircraft and pilot dynamics required by the system, and the pilot interface. Then results of a ight simulator evaluation with airline pilots are detailed. In 6 of 72 simulator runs, pilots were not able to establish a stable ight path on localizer and glideslope, suggesting a need for cockpit aids. However, results also suggest that, to be operationally feasible, such an aid must be capable of suggesting safe trajectories to the pilot; an aid that only veri ed plans entered by the pilot was found to have signi cantly detrimental effects on performance and pilot workload. Results also highlight that the trajectories suggested by the aid must capture the context of the emergency; for example, in some emergencies pilots were willing to violate ight envelope limits to reduce time in ight, in other emergencies the opposite was found. IntroductionR ESPONSIBILITY for the safe completion of a ight rests primarily with the pilot in command. During emergencies onboard air transport aircraft, this responsibility can be demanding, due to the large number of tasks to which the pilot must attend, including detecting and resolving failures in aircraft systems; continuing to monitor aircraft system health; coordinating with cabin crew, airline dispatchers, and air traf c control; controlling the aircraft; and deciding on (and then following) a course of action that will result in a safe landing. This inherent dif culty is compounded by a signi cant number of stressors, including physical danger, an uncomfortable physical environment (heat, smoke, noise, etc.), an overwhelming amount of information to consider, and the need to make decisionsin a short periodof time. In addition,the aircraftmay have degraded performance and handling qualities, limiting the extent to which the pilot's past experience is relevant to the present problem.The objectives of this research were to investigate how pilots generateand then follow a four-dimensionaltrajectoryto the runway threshold during emergencies and to examine the functions needed in pilot aids for these tasks. This paper rst presentsrelevantresearch from a number of domains, highlighting the important aspects of these tasks, pilots' needs in cockpit aids, and availabletechnologies. Then, the design of a prototype aid is described. The results of a ight simulator evaluation with airline pilots are detailed.The paper concludes with a discussion of pilot performance at these tasks and design recommendations for future cockpit systems.

Assessment of ECG and respiration recordings from simulated emergency landings of ultra light aircraft

“…Modern algorithms are usually based on genetic principles 22,23 . Some approaches that were developed tested trajectory design for emergency landings specifically in 20,24 .…”

Optimization of Flexible Approach Trajectories Using a Genetic Algorithm

“…27 This approach was also taken in other applications. 28,29 Other reasons for using segments include the facts that segments are a common way of describing trajectories both in air traffic control and in the flight management system and that an efficient communication or negotiation process between controllers and pilots can be supported.…”

Effects of modes of cockpit automation on pilot performance and workload in a next generation flight concept of operation

“…Previous research (Chen & Pritchett, 2001; Kaber et al, 2009; Kalambi et al, 2007; Wright, Kaber, & Endsley, 2003) has investigated the effect of various forms of cockpit automation on pilot performance and workload in replanning scenarios. The use of high‐level automation for flight path control (e.g., the FMS) has been found to support superior piloting performance in advance of critical events, such as an approach revision (Chen & Pritchett, 2001; Wright et al, 2003). However, the use of low‐level automation, engaging pilots to a greater extent in aircraft control loops, has been found to support superior pilot situation awareness and performance in dealing with an event, like an approach revision.…”

“…However, the use of low‐level automation, engaging pilots to a greater extent in aircraft control loops, has been found to support superior pilot situation awareness and performance in dealing with an event, like an approach revision. Related to this, Chen and Pritchett (2001) investigated the effect of different levels of automation on generating and predicting a detailed trajectory to the runway threshold in a simulated emergency situation. Under the highest level of automation, pilots exhibited automation bias (i.e., tending to follow the automatically generated trajectory because of high workload and time pressure), even when the tool generated erroneous trajectories (Chen & Pritchett, 2001).…”

“…Related to this, Chen and Pritchett (2001) investigated the effect of different levels of automation on generating and predicting a detailed trajectory to the runway threshold in a simulated emergency situation. Under the highest level of automation, pilots exhibited automation bias (i.e., tending to follow the automatically generated trajectory because of high workload and time pressure), even when the tool generated erroneous trajectories (Chen & Pritchett, 2001). Similarly, Wright et al (2003) observed that pilots using high‐level automation placed less emphasis on the accuracy of flight planning in advance of critical events and greater reliance on data stored in the FMS and the automation capability.…”