Do we really need pilots in aircraft?

Any flight you walk away from is a good one, right?

Sure, there are general expectations about relative comfort and convenience, timely departures and arrivals and a reasonable fare, but fundamentally, we want to get to the other end of the flight in one piece, ready to continue our onward journey.

The following discussion about having one or more pilots physically in and controlling the aircraft is presented through that lens.

The case for 2+ pilots

At the moment, large airliners need at least 2 pilots in the cockpit – on long haul flights there will be supplementary pilots as well, to allow the primary crew to get adequate rest during the flight.

There is growing interest from some airlines in having the International Civil Aviation Organization (ICAO) review international standards, with a view to allowing large airliners to be operated by a single pilot.

The benefit of this change would be reduced operating costs, and an easing of the global task of producing enough qualified pilots to operate all the aircraft that are expected to come on line in the coming years.

Various industry stakeholders have expressed a strong view that there must be a minimum of two pilots in the cockpit in larger passenger carrying aircraft.

The case for having two (or more) pilots in the cockpit of an aircraft is justifiably supported by high profile examples of where a computer would almost certainly not have been able to think its way through the situation and successfully resolve it.

An obvious example is where Captain Sullenberger and his offsider in US Airways Flight 1549 managed to ditch their aircraft on the Hudson River, both engines having been catastrophically damaged by the impact of a flock of Canada geese, with all on board walking away (albeit into boats).

Another example a bit closer to home is the QF 32 incident, in which damage was caused to the Qantas A380’s onboard systems by a catastrophic failure of one of the engines. The crew on the flight deck, with input from passengers onboard and external agencies, took about 50 minutes to complete all the initial procedures associated with the electronic centralised aircraft monitor (ECAM) messages, then formed and successfully executed a plan to recover the aircraft. While flying the approach to land, the autopilot disconnected twice in response to preset functions within the autopilot system relating to the aircraft’s angle of attack, so the Captain had to disconnect the autopilot and hand fly the final approach to land. The direct involvement of the five flight crew in the cockpit during this event almost certainly saved the 440 passengers and 24 cabin crew on board.

When things go wrong

Now for the counter argument.

It is an uncomfortable truth that, after proper investigation, a significant proportion of fatal aircraft accidents have had individual actions identified as a contributing factor.

During my time in the Air Force, I lost friends in aircraft accidents in which the individual actions of the pilot(s) or other crew member were contributing safety factors. Thankfully, the introduction of the Aviation Risk Management program in the early 2000s, along with a cultural shift in thinking, led to a significant reduction in fatal military aircraft accidents in Australia.

There are numerous examples of fatal passenger carrying aircraft accidents for which investigation reports note various individual actions as contributing safety factors. Think disorientation, loss of situational awareness, incorrect identification of the problem (with resulting incorrect solution), mishandling, pilot inattention, incorrect application of procedures, poor communication, descent below safety height / glideslope, unstable approaches, loss of separation, incorrect aircraft configuration, and more.

The Tenerife disaster is a case in point. On March 27, 1977, two B747 aircraft – KLM Flight 4805 and Pan AM Flight 1736 – collided on the runway at what is now Tenerife North Airport on the Spanish Island of Tenerife, with 583 fatalities. According to the final investigation report, the fundamental cause of the accident was that the commander of the KLM aircraft took off without authorisation, did not obey the air traffic control tower requirement to standby for take-off, the take-off was not interrupted when the Pan Am aircraft said that it was still on the runway, and when asked by the flight engineer whether the Pan Am aircraft had already left the runway, the Captain of the KLM aircraft answered with a resounding affirmation.

The report is available here:

https://www.mitma.es/organos-colegiados/ciaiac/publicaciones/informes-relevantes/a-1021977-y-a-1031977-accidente-ocurrido-el-27-de-marzo-de-1977/22-conclusiones

In Australia, on 7 May 2005, a Metro 23 with 2 pilots and 13 passengers impacted terrain in the Iron Range National Park while conducting a satellite-based non-precision approach to runway 12 at Lockhart River Airport. The Australian Transport Safety Bureau (ATSB) investigation report notes various contributing factors relating to occurrence events and individual actions. It also notes that the accident was almost certainly the result of controlled flight into terrain, that is, an airworthy aircraft under the control of the flight crew was flown unintentionally into terrain, probably with no prior awareness by the crew of the aircraft's proximity to terrain.?There are other contributing factors relating to local conditions, the aircraft operator’s processes and Civil Aviation Safety Authority’s processes.

The ATSB report is available here: https://www.atsb.gov.au/publications/investigation_reports/2005/aair/aair200501977

Aviation safety keeps improving

The good news is that aviation is getting safer all the time. Part of the reason for this is industry-wide programs to improve organisation and pilot / crew decision-making – think threat and error management, human factors programs and safety / risk management systems. Industry-wide education programs such as approach and landing accident reduction and runway incursion awareness have demonstrably reduced the number of safety occurrences.

Another significant contributor to improved safety outcomes is the ongoing development of technology.

The De Havilland DHC-4 Caribou, a twin radial STOL aircraft, didn’t have an autopilot - someone always had to physically control the aircraft. There were many times that two pilots, plus the flight engineer, were essential to the safe conduct of the flight. In terms of technology advancements, in the early 1990s, early model Trimble hand-held GPS were attached onto the dashboard with Velcro, as a supplementary aid to the old clock to map to ground method of tactical navigation.

Air Force Navigators became Weapon Systems Officers as their job evolved from managing (often coercing) relatively unreliable navigation and other aircraft systems to operating a complex suite of mission systems.

The classic B747-200/300s Qantas and many other airlines operated had a flight engineer’s panel with hundreds of switches and dials to manage fuel supply, weight and balance and other systems. These classic B747s are few and far between now – the flight engineer and their panel were replaced by code in an onboard computer in more recent versions of the type.

Boeing is delivering its final B747 this week, as advances in technology make it possible for twin-engine jets to replicate the Jumbo’s range and capacity at lower cost.

In the early 2000s, Qantas introduced instrument approaches that flew a curved path – they were called RNP-AR at the time. These approaches relied on more precise navigation accuracy enabled by improvements to onboard systems, plus training for the pilots and special regulatory approval for the operator. They made it possible to land and take-off at some airports during weather conditions that would otherwise have prevented safe aircraft operations, and improved operational efficiency by reducing track miles on approach and departure.

Machines and systems are getting more reliable

The increasing reliability and effectiveness of automation technology has changed the pilot’s job from doing to monitoring – a job humans are generally not good at.

100 years or so after the first commercial passengers were carried by aircraft, it could be argued that the pilot has become the weakest link in what is, in modern aircraft, a highly reliable system.

Modern aircraft are designed to be reliable – we’re talking a catastrophic failure condition occurring once in a billion flight hours.

In this context, the recent inflight shutdown of an engine on QF144, a B737-800 flying from Auckland to Sydney, is interesting to consider. Could it have been handled by a single pilot in the cockpit with external support, or even one or two pilots in the loop but somewhere else?

This brings us back to the question of how reliable are pilots and do we really need them in the cockpits of passenger aircraft, whether we’re talking one, two or many?

Autonomous (human in the loop) passenger carrying operations will be here soon

The myriad of reasons for not progressing to single pilot airline operations are already informing the development of autonomous (human in the loop) operations.

Large uncrewed aircraft have been operated/monitored by pilots located many thousands of nautical miles away, for decades now, mostly in military applications.

I expect it won’t be too long before we see large freighter aircraft without pilots in the cockpit - they might be some of the current fleet, retro-fitted with the necessary equipment to be operated remotely, or new types designed from scratch.

Electric vertical take-off and landing aircraft developer Wisk is marketing its 4-passenger advanced air mobility aircraft as ‘autonomous (self-flying) with human oversight’. That means no pilot onboard. The concept of operations for these aircraft is to have a multi-vehicle supervisor, who will manage multiple aircraft remotely (up to three initially, then many). Wisk is expecting to be operating these aircraft by the end of the decade (2030).

I think we’ll eventually get used to the idea of not having pilots physically in the cockpits of passenger aircraft. We’ll just need to make sure that someone is still able to turn off or override the computers and take over when things go wrong.

Having only one pilot onboard large passenger aircraft is probably not such a big leap, provided the system is designed accordingly.?