Aviation Week Op-Ed: Embry-Riddle President Puts Forward New Approach to Resilient Performance in Aviation Safety

Captioned video from simulator sessions that are being analyzed by an Embry-Riddle and NASA team to identify positive pilot performance
Captioned video from simulator sessions that are being analyzed by an Embry-Riddle and NASA team to identify positive pilot performance. (Image: Embry-Riddle)

In his latest Aviation Week essay, Embry‑Riddle Aeronautical University President P. Barry Butler, Ph.D., explores how researchers and airlines are studying positive performance to improve aviation safety. Through the Hunt Library, the Eagle community can log on to ERNIE to access the op-ed. Alternatively, subscribers to Aviation Week can log on to the publication’s website to access the essay, which is also provided below.

By P. Barry Butler

While aviation tends to focus on learning from mistakes — and rightly so — the potential to learn from positive outcomes often goes unrecognized. 

Dr. Jon Holbrook, a cognitive scientist specializing in aviation at NASA’s Langley Research Center, says that pilot intervention has “kept millions of flights safe.”

Pilots intervene to manage aircraft malfunctions on a fifth of normal flights, according to a 2013 joint report by a team that analyzed safety and operational data from the Federal Aviation Administration’s Line Operations Safety Assessment program.  

Holbrook, by extrapolating this finding to worldwide jet data on 266 million flights between 2007 and 2016, estimated that pilots kept flights safe from aircraft malfunctions nearly 49 million times during that period. This means, according to Holbrook, that crews intervened 157,000 times for every time that human error was implicated in an accident during that same ten years.

The reasons for that success are harder to measure. Investigative resources to date have focused on negative outcomes, for good reason. But as a result, the system is not designed to capture the routine yet unexpected events that pilots manage safely.

Research has also shown that pilots are not likely to reflect on their day-to-day challenges — from sick passengers to traffic conflicts — and how their planning, adaptation and coordination help them overcome these challenges. That is why a team of researchers from Embry-Riddle and NASA is working to find ways to capture systematically the practices that keep the U.S. aviation system — the largest and most complex in the world — so safe.

A key objective is collecting and analyzing data on specific pilot actions taken to manage operational difficulties. Researchers have already begun to address this knowledge gap.

First, they are using automatic speech recognition to analyze audio and video from simulator sessions in which pilots navigate real-world challenges, including heavy traffic, weather diversions, shifting winds and wake turbulence. The recordings also capture video of the pilots reviewing their performances as they narrate play-by-plays of their flights.

The Embry-Riddle and NASA team is now applying coding and natural language processing to identify resilient performance and evaluate the impact of the pilot debriefing sessions.

“When we analyze those data, we are looking for evidence of resilient performance — what did the crews do that helped them anticipate problems, plan for possibilities and coordinate their response?” explains Dr. Kristy Kiernan, associate director of the Boeing Center for Aviation and Aerospace Safety at Embry‑Riddle and the project’s lead investigator. Kiernan also serves as program coordinator for the Master of Science in Aviation Safety at Embry-Riddle’s Worldwide Campus and as an associate professor in the Department of Aeronautics.

Second, the team is developing a large language model to analyze National Transportation Safety Board accident reports and dockets for evidence of resilient performance, with the goal of maximizing lessons learned from these unfortunate events.

Major airlines are working with their pilots on various debriefing methods. Both American Airlines and Southwest Airlines have dedicated teams and resources to programs that collect data on positive performance. Directors of these programs say they have seen results.

Capturing more granular information about operations on the 45,000 daily flights in the national airspace will require new tools. Fortunately, newer aircraft are collecting more data. Tools and techniques for analyzing it, including artificial intelligence, hold promise for “insights into routine performance that were not possible before,” Kiernan says.

Finally, aviation is entering a new era of automation and autonomous flight. An already congested airspace will need to accommodate a growing number of new entrants, including more drones and breakthrough categories such as advanced air mobility aircraft. Flight controls are evolving, and pilots will need to learn new technologies and flight capabilities.

For now, however, it is in the “human element where most of the resilience resides,” Kiernan says, and that resilience is what needs to be captured to reinforce safety in this emerging system. 

Aviation has always benefited from a culture of continuous learning, and our industry must learn from all operations for this positive, resilience-based approach to work. We’ve created the world’s safest form of transportation through obsessively reporting our errors. Imagine what our industry could achieve if we examined our successes with equal rigor.