Rise of the Robots: Automation in Commercial Aviation

Several years ago, I went on a first date with a pilot for a major international airline. We visited an Air New Zealand exhibit which discussed the history of Air New Zealand and New Zealand aviation, and also proposed what the future of aviation could hold. On a large whiteboard wall, guests to the exhibit were asked to write their comments about futuristic aviation concepts, including the concept of ‘robot pilots’.

My date scoffed that anyone supportive of the ludicrous concept of robot pilots was obviously uninformed, and wiped a few of the comments off the wall. I laughed but could tell it had touched a raw nerve. After all, I knew that automation already played a significant part in his job. He often joked that being a pilot consisted of a lot of procedural rote learning and memorization. In his words, this resulted in a job “pressing buttons and staring out the window.” After all, with 57% of jobs changing as a result of automation, the increased use of algorithms or machines, pilots are not immune to technological changes in the future of work.

A recent report compiled by the New Zealand Airline Pilots Association stated that pilots needed to be “well educated (but not brilliant).” In short, pilots should be smart but it’s not essential for them to be that smart. After all, pilots are just people. Some have mental health issues, despite the various measures put in place to prevent them from operating aircraft. Some act in an inappropriate and emotionally-charged manner. Potentially due to increased automation, some are forgetting how to do their jobs. Both robots and humans are fallible. This has begged the question – what if we could move to a future where automation was advanced enough to outweigh the risk of that pesky potential of human error, or at least to reduce the ‘human’ element to flying?

Remember When There Were Five People on the Flight Deck?

If you flew on a commercial airliner in the 1950s, you could have expected a crew of five airmen operating your aircraft – two pilots, a flight engineer, a radio operator, and a navigator. By the time the 1960s rolled around, the radio operator and navigator were gone, and flight engineers were removed during the 1990s. It’s not unusual to have a crew of only two pilots – a Captain, and a First Officer – on flights up to around eight hours, where standard risk management regulation requires a third operating crew member to be on the plane. As you can imagine, this can result in a fair bit of time when there is only one pilot technically operating the aircraft – or at least on the flight deck – in light of inevitable bathroom breaks, leg-stretching walks, and controlled rest. It’s not such a stretch to think that, within the next ten years, airlines will want to think how they can reduce numbers further – and the possibility of a single-pilot cockpit is called into question.

The Robots Already Rule the Air

Automation already plays a massive part in aviation. The decline in flight deck crew numbers? It’s due, in a large part, to the rise of automated technologies that have eliminated many of the routine pilot functions of yesteryear. Most modern aircraft are controlled mainly by computers, with automatic flight management systems. The role of the pilot has been altered considerably, given the advancement and accuracy of these computerized systems. Pilots can, of course, take over the controls – but systems and processes are in place to allow aircraft to operate under considerable automation, and in fact, this has become the accepted norm for commercial aviation.

Why Would Airlines Entertain the Idea of the Single-Pilot Cockpit?

UBS has claimed that around USD $15 billion in savings could be achieved by airlines, annually, by using a single pilot cockpit. This ambitious figure does hint at a growing issue amongst airline operators – the cost of keeping all of those pilots flying.

As China continues to attract pilots by offering large, tax-free salaries, it’s becoming more difficult for a number of airlines to attract and retain pilots. The 2019 Boeing Pilot & Technician Outlook estimates that 804,000 new pilots will be needed to fly the projected aviation fleet over the next 20 years. That’s around an additional 19,000 pilots each year. Of course, these projections don’t yet take into account the potential impacts of single-pilot commercial airliners – which could be a massive shake-up for the industry.

What About Stick-And-Rudder Flying Skills?

Late last year, my article The Pilot Paradox discussed the concept of the Multi-Crew Pilot Licence (MPL), designed to get pilots operational faster than on a traditional Commercial Pilot Licence (CPL) route. Premised on the concept that pilots intending to fly in a multi-crew cockpit environment should learn how to work in that environment from day one, it requires much less time spent manually operating a real aircraft, and more time in a simulator environment.

Apart from the obvious issue of a ‘multi-crew’ license in a world of single-pilot cockpits, the fact remains that all aviation training faces greater levels of automation than ever before. There is a greater emphasis on the utilization of emergent technologies at all levels of aviation training. While advancements in technology have, undoubtedly, made flying safer, there is a lingering question – how much manual handling should airline pilots continue to be taught?

When Automation Goes Awry

In October 2018, Lion Air flight 610 crashed, killing 189 people. In March 2019 Ethiopian Airlines flight 302 crashed, killing 139 people. Both planes met their end in eerily similar circumstances, with an erroneous sensor—a sensor designed as a new safety measure aimed at preventing pilots from losing control—pushing the aircraft into a steep dive. The Maneuvering Characteristics Augmentation System (MCAS) was put in place to account for design changes in the new 737 MAX, helping to stop the nose of the plane from pitching up due to the altered placement of its engines. Because of the very specific operating parameters under which Boeing expected it to activate, details on its usage were limited.

In November 2018, Boeing admitted that it had failed to include information on the MCAS in its training manuals for 737 MAX aircraft. However, even after manuals were updated, it was not enough to prevent the Ethopian Airlines crash in March 2019. In that case, it appeared to be a mixture of capability, training and technological design flaws that brought down the plane. What has resulted, therefore, is a mixed bag of questions: how safe are some of the ‘new’ automated technologies on aircraft? Are pilots gaining sufficient training to understand these systems, and to know how to cope when they fail?

Even though Boeing needs to respond to the issues surrounding the 737 MAX (an obvious one being the functionality of warning lights), questions are also being asked as to whether airlines are cutting corners in training. Some aviation industry figures argue that a shift in focus for pilot training has led to a decline in basic manual flying skills, which at a minimum would require the ability to manipulate the stick, rudder, and throttle to keep a plane at the correct speed, pitch, and altitude. In a 2017 video, Ed Wilson, Chief Pilot for Boeing, bragged that in just 2 1/2 hours of computer-based training, currently certified 737 pilots could be ready to transition to the 737 MAX. Would any airline believe this is sufficient?

When renowned pilot Chesley Sullenberger is weighing in on the question of automation and training, we should probably pay attention. The man who safely landed the ill-fated US Airways flight 1549 stated that “Pilots are not being told or taught everything they need to know about their airplanes.” His argument that automation requires an entirely new type of training holds weight, given the potential for overreliance on automated systems. He emphasized enhanced simulator training, as opposed to the current shift towards computer-based training. In June 2019, he stressed that it was unlikely that any airline pilot in the U.S. would have been taught how to deal with the warnings that arose leading up to the 737 MAX accidents. This is an important reflection, given some of the discussions about the potential poor quality of training of the airlines involved.

Aviation Is Safer Today Than It Has Ever Been. But What Can Improve?

Despite these accidents, commercial aviation is now safer than it has ever been. Between 2010 and 2019, just one person died on a U.S. commercial airliner. Around 100 people die every day in the U.S. from injuries sustained in car crashes. Aviation is safe. However, aviation is also big business, and there is room for improvement. Whether these improvements address financial performance or managing the risk of human error, there are plenty of motivators to explore automation in aviation.

In 2017, Airbus started trialing Vahana, self-piloted vehicles with a range of 50 miles. Airbus stated that they were aiming to take them—commercially—to the skies in the next ten years. The program saw its first test flight in February 2018, and by March 2019, had undertaken 50 flights—now moving towards the second iteration of its prototype.

At a briefing ahead of the Paris Air Show in 2017, Boeing announced it was investigating technologies for self-flying aircraft, noting that current aircraft can take off, cruise and land using automation. In November 2017, it acquired Aurora Flight Sciences, completing its first test flight of the Boeing NeXt in January 2019.

There are other ways that automation and A.I. might make flying safer. The use of biometric data, for one, is ensuring that greater oversight is being gained over those that take to the sky. An April 2019 conference on Vienna was held to understand how biometrics can support the UN Security Council Resolution 2396, a milestone decision in efforts to detect and prevent the movement of terrorists.

A Single Pilot Cockpit—the Game Changer?

In light of pilot shortages and increased demand for air travel, advancements in single-pilot cockpits could foment radical change. While rumors of the proposed 797 release at the Paris Air Show have not come to fruition, murmurings continue around single-pilot cockpits.

Jefferies Financial Group analysts have claimed that the 797 could come with a total redesign of the cockpit environment, potentially with one pilot flying the aircraft—and a second pilot monitoring the activities of several aircraft from the ground. While this could have been the big reveal expected at Paris, Boeing has plenty of other issues on its hands. It issued a statement confirming that its new midsize airplane will be based upon a twin-pilot premise. It will be a little longer before we can understand just how the major aircraft producers envisage the introduction of a single-pilot cockpit.

Will the Public Accept a Single Pilot?

Regardless of how you approach the question—humanity has a way to go before we are likely to accept autonomous, or near-autonomous, aircraft. Every incident of supposed ‘human-error’ begs us to question how we choose, and train, those allowed to shuttle us around the world. Despite this, most people simply aren’t in the headspace to accept an autonomous plane. In 2017, over 50% of people questioned by UBS stated they wouldn’t fly in an aircraft without a human pilot – even if fares were cheaper.

While trials for self-flying taxis continue, the fundamental shift required in our thinking to accept this type of approach to aviation is a much more difficult problem to overcome. Boeing’s Vice President of Research and Technology, Charles Toups, has noted that single-pilot technologies will likely be introduced into cargo operations first, and it would take several decades before we will see passenger operations being undertaken in single-pilot aircraft.

We might be averse to having a single-pilot cockpit on a passenger jet, but are we that opposed to having our cargo delivered by them and sharing our airspace? Time will tell, but at the pace that technology is progressing, it might happen sooner than you think.

 

[Featured Image: iStock]