“It’s all on autopilot, right?”

Autonomy versus automation, a perennial discussion. But the nature of human intervention in flight is much more subtle than simply the mechanics of flying. The autopilot was selected as we climbed out of Heathrow. The blue sky beckoned, the south coast almost unnaturally stark on the clearest of summer days. A heavy Airbus A321 needs a little bit of attention on a turning departure so we had planned a pause in the flap retraction sequence until the heading change was complete. It afforded us a better turn radius. A bit more margin over the lower end of the speed range. A bit more resilience should we bounce through a bit of wake.

The autopilot stayed engaged. The First Officer (FO) selected modes. Finessing the performance of the machine. Stabilising or moderating the climb rate. Depending on what best suited the environment at that moment. Hundreds of tiny interventions. Direct to here. Stay on this heading. A nudge around the weather. How about that step climb? The jet can do it but how about this choppy turbulence? Sensible?

Sliding along the Croatian coast the colours of the sky bled out through a deep orange and in to darkness. Tweaking the lighting knobs to drop the ambient light the first stars peeked through the blackness as the carpet of the night unrolled. We planned out our strategy for picking our way through Cypriot and Israeli airspace. The timing plan. The carefully coordinated activities. Who is doing what and when? Who is doing the flying? What are our go-around and fuel strategies? Building some area situational awareness. We have a feel for what is where, what the weather is like, what the challenges might be in the couple of hundred miles surrounding us.

Beyond the inch or so of windscreen lays a frigid environment so hostile that we only ever hold it at bay. Gently but firmly so. The trick and illusion of flight. The last of the wine is poured in the business cabin behind us. The dessert and cheese is cleared away. The click of seatbelts. The chime of the overhead signs. Latches snap down. Tel Aviv change the altitude restriction at the Jordanian border. “Amman want you at one six thousand. Descend and maintain one six thousand”. I dial 16,000 into the FCU and we to and fro with some standard calls.

IMG_3268Strange. That is 5,000 feet higher than we were expecting. Maybe it is a traffic issue? Some departing flights? Military activity? This part of the world is simmering at the moment. We discuss the implications. I update my ‘energy plan’. We will be significantly above a workable profile. Losing either speed or height in a jet can be a challenge and I have some options. First, I can trade some kinetic energy for the enforced potential energy; we might be high, but we can be slow. Our total energy state is a sum of both. I spin the speed knob back to about 220 knots. The airflow and engine noise is low. Like we are gliding into the velvety night above the Middle East.

We plan a potential scenario where we are suddenly cleared to a much lower level. Perhaps if the assumed and invisible traffic has moved away. I discuss the best autopilot mode, how I would wind the speed up and use the speed brake to wipe off energy. I am currently flying as efficiently as possible. Preserving fuel and with it decision-making capacity. The next descent plan is almost opposite; It is an exercise in flying inefficiently to fit back in with potential air traffic control requests.

I scan the fuel quantity. 3200 kilos left of our original 17,700. I did a quick sum to convert it into a time range at our expected burn, then looked at the clock. The Jordanian controller on the next frequency sounded fractious. He was working at his maximum capacity. Something had occurred at his airport that meant it was shut to all traffic. This was not what we needed to hear. I looked across at the FO as we were sent to a holding fix. Indeterminate delay. The controller actually used the term ‘indefinate’.

Fly first. The rapid fire instructions and information meant lots of changes to the automation. We are going where he wants us to. Good modes meaning that the flight path is safe. Holding fix is programmed. A steady pace of interaction with the automation. A pause to check that it has had the desired effect. On to the next task. Done. Another glance at each other: what is going on? Can we stay here or do we need to start thinking about going elsewhere?

Aircraft spiral around the holding fix above and below us. The controller is working so hard that he is making a few slip ups. We resolve to check his instructions carefully between ourselves to goalie for him and to avoid getting so saturated with our own tasks that we stop keeping our ears alert to his instructions. We remind ourselves again of the airport elevation and local minimum altitude.

We need more information on which to build a plan. The objective is to never be left only holding one card. And to always keep an ace. We need to work with the known and progress logically through the options. Safe is the word, legal too. Our approach to the problem needs to measured. Hope, as my old boss used to say, is a wonderful virtue but a terrible strategy.

“There is an air traffic control critical incident ongoing”. Dramatic. It set my mind racing. A fire in the tower? Some form of terrorist attack? Maybe the approach systems or lighting has failed? Turning close in above the airport dispelled these last two theories: the lights were all on, the navigation aids were all working. The clock was ticking as the fuel ran through the pipes. Politically, Tel Aviv was not top of my options list. Marka airport was almost next door. But what if it is affected by the same issues?

The controller was unable to provide an expected approach time. Without either that or a bit more information about the problem we were running on faith. We had the outline but to keep the passengers safe I needed a bit more detail. “Is the incident affecting an aircraft or the ATC facilities?”, “Are approaches available at both Queen Alia and Marka airports?”. The answers were quick and simple: the incident was something to do with an aircraft and both airports had full approach facilities.

The game changed. Another glance at each other. Our ace was Marka. We could sit in the hold and burn down a much lower fuel figure. If the main airport did not come back on line then we could bank right and make an approach to the secondary airport. We had a clear line in the sand, once we got there we had options. The FO rippled through the pages of the QRH, checking landing performance and we both set up charts in the iPad. The basic planning for Marka meant that if we needed to go there all we had to concentrate on was flying the aircraft and we could keep the chat to a minimum.

Some runway exits were closed as a result of the incident so the final task was to perform a shorter landing than normal. The FO clicked out the autopilot at 900 feet. The disconnect warning sounded and ceased. She guided the Airbus into the touchdown zone, wasted no time pressing the main gear gently on to the concrete and ‘lowering the nose. The nose wheel touched as the braking and reverse thrust started to bite.

I get told that pilots no longer ‘fly’ airliners; a fact that only stands up to scrutiny if your definition of flying is solely moving the controls. To keep my customers safe and sound and to guide them to where they want to go you need my brain just as much as you need my hands. We sat in the flight deck and I scribbled in the aircraft log book to put the jet to bed. We discussed the events of the last 30 or 40 minutes. Could we have recognised the issue earlier? Were there better ways of getting the information that we needed? Were we happy with the outcome?

The flight deck door was open, customers were filing off. I watched them moving up the glass walkway to the terminal. Families, children, business men and women, tourists. Success for me was their lack of awareness of the difficulties of the last part of the flight. I would never have placed them in a situation where I was relying on hope. “Nice landing, captain!” one of them called. “You can thank the First Officer for that” I replied, pointing across the flight deck.

Autopilot does not mean autonomous. Not flying by hand does not mean the skill is not needed. It will be a long time before the aviation environment truly supports complete autonomy. There will be pilots for a few years yet. Bag zipped up. Jacket on, hat, sunglasses, iPad, phone. The last few switches thrown. Handed the jet to the engineers. “Have a great evening”.

20 thoughts on ““It’s all on autopilot, right?”

  1. Amazing story! I love it! It kept me reading right through. My type of flying in much more hands on and lacks the political issues you face. But I see many common threads. Weather, runaway conditions, fuel reserves, and always having a solid out or ace to play are keys to my flying as well. Thanks for sharing!

  2. Nice article MMS!
    When people say to me “all you do is press buttons” I reply ” yes in the same way that all JK Rowling did to write the Harry Potter stories was press keys on a keyboard”!

      • Hello Mark,
        Great article – I know it’s a while since you penned this but I’d like to include it as part of the reading list for a Threat & Error Management Module I’m involved in teaching (I’m ex BA and a current CRMI). I can’t see your full name to be able to credit you though – can you let me know? You perfectly capture the subtle threats that are more likely to crop up, and some great practical management of these – it’s not all about engines dropping off is it?
        Be good to hear from you if this thread is still working,
        Best wishes
        Sue

  3. Interesting article! I do wonder however what the perspective of air traffic control or of air traffic management would be for that flight? Could a more holistic understanding of the challenges of that flight have led to a more efficient out come which required fewer decisions to be made, fewer machine – human, human – machine and human – human interfaces which could have made the flight even safer? Could the issue at the destination have been prevented, could your route have been better managed with earlier knowledge and a wider view got you on the ground sooner? Yes it would require a super computer to manage this, but we already rely on supercomputers for our weather forecasting.
    As a fellow ATPL holder I would like to see the role of pilots continue but I can’t feeling that as machines and software get more reliable, the organic matter in the front becomes the ever weaker link in the chain, making autonomous flight inevitable and in fact more desirable for the passengers. Yes humans that design software and systems are also fallible but offline QA is much more robust than on line in flight quality assurance.
    Interesting topic!

    • Hey Ben, interesting point. We were, of course, responding only to the information that was passed to us. It turned out to be a disabled aircraft so as such, not an ATC flow management issue.

      It was a human problem with human solutions. I’m not sure what the algorithmic response would be but it would be interesting to find out.

      Thanks for reading,

      Mark

  4. Excellent and thought provoking. It strikes me that the skills I had to learn as a pilot I am now having to apply in my fancy new car. It is often hard to distinguish between the various modes and the similar-sounding ‘alarms’.
    It took us a while to learn the lessons of CRM and automation in the cockpit, I wonder how long before we have to go through a similar process in our cars?

Leave a reply to Ben Clarke Cancel reply