Concurrent Pilot Instrument Monitoring in the Automated Multi-Crew Airline Cockpit

Abstract

Pilot instrument monitoring has been described as "inadequate," "ineffective," and "insufficient" after multicrew aircraft accidents. Regulators have called for improved instrument monitoring by flight crews, but scientific knowledge in the area is scarce. Research has tended to investigate the monitoring of individual pilots when in the pilot-flying role; very little research has looked at crew monitoring, or that of the "monitoring-pilot" role despite it being half of the apparent problem. Eye-tracking data were collected from 17 properly constituted and current Boeing 737 crews operating in a full motion simulator. Each crew flew four realistic flight segments, with pilots swapping between the pilot-flying and pilot-monitoring roles, with and without the autopilot engaged. Analysis was performed on the 375 maneuvering-segments prior to localizer intercept. Autopilot engagement led to significantly less visual dwell time on the attitude director indicator (mean 212.8-47.8 s for the flying pilot and 58.5-39.8 s for the monitoring-pilot) and an associated increase on the horizontal situation indicator (18-52.5 s and 36.4-50.5 s). The flying-pilots' withdrawal of attention from the primary flight reference and increased attention to the primary navigational reference was paralleled rather than complemented by the monitoring-pilot, suggesting that monitoring vulnerabilities can be duplicated in the flight deck. Therefore it is possible that accident causes identified as "inadequate" or "insufficient" monitoring, are in fact a result of parallel monitoring.Jarvis SR. Concurrent pilot instrument monitoring in the automated multi-crew airline cockpit. Aerosp Med Hum Perform. 2017; 88(12):1100-1106.