Very Stressed and Rightly So

December 18, 2013 by Bruce Landsberg

757  cockpit (9)The two most-often heard statements in highly automated cockpits are said to be, “What’s it doing now?” and “Look, it’s doing it again!”

The Korean pilot in command of the Asiana Boeing 777 flight that crashed at San Francisco made a profound admission at the NTSB hearing last week. The captain, who was in training, found the visual approach “very stressful” and “difficult to perform.” The ILS was down for maintenance and the weather was perfect. He stalled the aircraft on short final, erroneously thinking that the autothrottles would save him in a bad case of mode confusion. Go figure.

Automation dependence, as it is now being called, is the growing realization that automatic everything in our cockpits is probably not the panacea that many thought it would be. This landmark accident may just help the industry and the regulators get their collective heads out of the automated sand. Seems like there is good reason for many of us to feel “very stressful.”

Come with me back to the days of yesteryear—December 1995. Recall when a perfectly functioning American Airlines Boeing 757 on a night approach into Cali, Colombia, was programmed to fly nearly 90 degrees off course? That’s not a good thing to do in the Andes, and a mountain intervened. The crew was confused about programming a complex flight management system even though they had a lot of experience.

American, which has flown into South America successfully for decades, decided it was time to reprogram their crews to: 1) remain proficient in basic flying skills, and 2) when confused by the automation—turn the blooming thing off and fly the aircraft. The technical term was to go down in levels of automation. Good idea!

Basic automation is very good and allows us to fly more precisely and with less fatigue. But when it takes more effort to learn and remember how to make the aircraft fly on its own than to just do it yourself, then we’ve outsmarted ourselves—or perhaps we’ve been outsmarted by those who somehow think that more automation is better.

All that glitters is not glass and servos. For the training academies and schools preparing the next generation of pilots, this is a modest suggestion to start with basics no matter how seductive the push for glass is. Not being able to fly a visual approach is a stunning indictment! For the manufacturers of airframes and avionics, look back at the spate of air carrier accidents where automation and complexity was at least a contributing factor. As I wrote more than a decade ago, “Simpler is safer, and if a massive amount of training and button pushing is required to operate the magic, then maybe we need a smarter magician. The lessons of Cali for pilots, avionics designers, ATC, and regulators alike are written in blood.”

Two other thoughts: Angle of attack is a critically important and basic flight function, and when a trainee is getting outside the parameters—even if he or she is a senior pilot—the instructor has the absolute obligation to go around! We could go on for days on this topic!!!

Looking Ahead: Let me wish you and your family a safe and healthy aviation new year and thank you for joining us. Special thanks to the donors who support the AOPA Foundation and help preserve our freedom to fly. The blog will stay in the hangar next week but will be airborne again after the first of the year. Until then—safe flights everyone!

Bruce Landsberg
Senior Safety Advisor, Air Safety Institute

ASI Online Safety Courses  |  ASI Safety Quiz  |  Support the AOPA Foundation

  • lindsay petre

    Apparently the lack of GA in some other countries means that airline pilots do not start out with the basics that we take for granted but are thrown directly into complex simulators.

  • Dan Winkelman

    Right on, Bruce. As both a pilot and aerospace systems engineer, this is precisely what I advocate. No matter how fancy the system, there needs to be an easy way to turn it off, revert to fundamental flight and nav instruments, and hand-fly the airplane. All pilots must maintain proficiency in hand flying. The “systems manager” approach advocated by some is going to result in more loss of life. When seconds count, you don’t have minutes to spend jerking around with the system. You need to be able to say “screw this, I’m flying” on a moment’s notice… and pilots should be trained in how to do exactly that. The avionics systems, likewise, need to recognize that they are there to *assist* the pilot, not do his job for him.

  • Russ Miller

    Having flown both the MD-11 & B-747-400, I’ve always said and taught my F/O’s ; ” Never forget that behind all the glass and magic is a stick and rudder airplane”!

  • David Consbruck

    We rarely see anyone “go around”. Frankly, I believe it may be the most powerful tool in the pilot’s landing bag. On the two or three commercial air carrier flights where, as a passenger, I experienced a go around, I felt relieved, safe, and proud of the pilot in command taking charge of the aircraft and the situation. As a student pilot and as flight instructor, I found the “go around” to be one of the most assuring and “in control” feelings. Remember those words, “Every landing is optional.” ? Well, on the few occasions when they are not, the accident chain has already started. I am going to add “go arounds” to every flight review I do in the future. I know it will build pilot proficiency and confidence. And, that will increase flight safety.

  • Bruce Landsberg


    Bravo!!!!! Go Arounds SHOULD be a part of every flight review.

    I have found over years of training that when I had an approach nailed, perfectly set up,the instructor would invariably say that a cow had wandered onto the runway – ergo – Go around.

    It’s a tradition I’ve carried on with my own students. It makes for a lot less hamburger!

  • Don

    All the Korean nuclear plants have been shut down because the maintenance documentation has been discovered to be about as legit as their pilot training. (google “Capt Tom Brown UAL”)
    Yes, the new automation is a challenge to those who don’t like to study and practice, but let’s not tar the whole effort with one aircrew who didn’t know what a stick and rudder airplane was, let alone remember that they were flying one.

  • JT

    One of the things I have seen as a captain is the newer pilots out of schools that were trained in glass cockpits have a very difficult time flying our old steam six packs. They have no scan and cannot hold altitude or course very well. The anxiety level in actual IFR is very high for them. Schools need to realize that there are many places that still use older equipment (AA’s Super 80s, commuters using Cessna Caravans, etc.) and train students to be comfortable on both.
    As for the Asiana captain, I find it unbelievable that a visual approach was stressful, especially in severe clear weather. Shooting an approach to minimums is stressful. That day the crew should have switched off the autopilot and hand flown it. Some pilots forget the simple pleasures of what flying really is.

  • Charley Brown

    How much was Asiana paying a guy who was incapable of performing a straight-in landing on a beautiful VFR day? What was his job description?

  • John Schreiber

    Besides the lack of basic flying skills, it is important to know how things work. With regard to automation, perhaps we need the flight engineer again, to manage the complex systems.

  • Duane

    It’s silly to blame automation for these accidents, which only deflects earned blame from the Pilot in Command. The sole responsible party for the operation of any aircraft is the PIC. If he or she is not in full command of the aircraft at all times, then he or she isn’t doing their sole job and fulfilling their legal responsibility.

    I had a discussion of this accident with my wife a couple weeks ago. I found it a bit difficult to communicate clearly with her, because the distinction between “flying the aircraft” and letting certain tasks be performed routinely by auto controls isn’t easy to explain, especially to non-pilots.

    My wife was of the view that the excessive reliance on automation was at fault, whereas I put the fault solely on the PIC. She argued that it’s only human nature to rely excessively on the crutch of cockpit automation, and thus to become mentally lazy. I argued that pilots, especially airliner pilots entrusted with the lives of hundreds of souls, simply have no other job but to fly the airplane, all the time. The PIC has no business “relying” on the airplane to fly itself.

    I’m not sure that I convinced her. But even other pilots seem to make the same mistake when they blame excessive automation for accidents like this. It’s not a matter of excessive automation – it’s a matter of insufficient control. It’s a distinction that makes a big difference.

    It is absolutely necessary that pilots of automated cockpit aircraft receive recurrent training, both in simulators and real aircraft, that involves actual manipulation of the flight controls and throttles, in both standard maneuvers such as visual approaches and instrument approaches, as well as in various simulated failures of the auto controls.

    It is also reasonable for airlines and other organizations to require that their pilots perform some minimum number of hand-flown flight segments and maneuvers per total number of missions. Perhaps the FARs should also address this need.

  • Hector

    I agree with most of the comment, but when you had been flying a plane for more then 10 or so hours, even with the perfect weather you start feeling tired, and don’t forget that we had to do the ordinary stuff that requires every day.
    I think that the airlines have to start thinking very seriously about the REST OF THE CREW.

  • Austin Kalb

    I have watched this video on automation dependency many times. It is absolutely one of my favorites!

  • Michael Gibbs

    Even if this pilot was taught to be a systems manager rather than a pilot, he didn’t even do that well. How long would you sit and stare at that decaying airspeed indication before intervening?

  • Alan Edgren

    Back in the dinosaur age (1968-69), I was learning to fly in a Cessna 140. After a few hours of dual, I asked my instructor (man, could he fly an airplane!) what the mysterious rectangular box low on the panel was. “That’s a radio. Don’t touch it” was his response. I bet I had thirty hours in that airplane before he turned on the radio and showed me how to use it. I learned how to fly first.
    As an airline pilot, I flew an Airbus 321 into KSFO many times. I would routinely remind my copilot and myself when flying the FMS Bridge Visual that it’s just an airplane and we can shut off the fancy stuff and use the flight controls and throttles the same as in a Cessna 140. Airbus 321, Boeing 777, Cessna 140. All just airplanes requiring a PILOT.

  • Fred Weems

    I especially liked the part about an AOA indicator. The Aardvarks I flew back in the stone age had AOA calibrated in degrees, and I have really missed that info ever since I left that system. So simple and cheap to implement, so rarely even considered.
    Given the potential inaccuracies inherent just in the weight and balance routines in any given part of the industry, an AOA indicator could be a lifesaver in so many situations, like when I belatedly discovered an error of over 10,000lbs on a weight slip one time.
    I am happily retired from 40 years of flying now, but for those of you still working hard to keep the blue side up, I hope Santa puts an AOA indicator into your well earned pile of gifts.

  • Avi Weiss

    In the original “Karate Kid”, Mr. Miyagi admonishes Daniel about commitment to Karate via a analogy: “You walk on left side of road, ok; You walk on right side; ok. You walk middle of road, pretty soon, squish, just like grape”. So too, with current aircraft systems automation and flight control, we find ourselves as an industry in the middle of the road, with a plethora of flight automation tools but having a bifurcated and somewhat arbitrary operational mindset between human and automation control, driven by inconsistent user interfaces, poor lines of human-to-automation demarcation and modes for control procedures, and a legislative, financial, and to some degree, emotional inability to fully commit one way or the other.

    Leaving aside the issues of basic airmanship skills and crew responsibilities, while there has been much electrons spun to debate the relative merits of automation and its complexity, and the evolving role pilots will play in the face of increasing automation, I feel the discussion has frequently focused on trying to “assess the un-assessable”, with resultant arguments about whether “more automation is good or bad” based on conflation of emotional issues with abstract decision making and stick and rudder skills. To wit, Bruce’s comment of that “perhaps we’ve been outsmarted by those who somehow think more automation is better” may be making a valid point about “more automation”, but I would counter that a more meaningful posit would be that “BETTER automation is better”, thus moving the focus of discussion from “quantity” to a more useful one of “quality”.

    When the current automated systems are fully functional and being operated AS DESIGNED by PROPERLY trained and PROFICIENT pilots, “automation” will consistently fly an aircraft more accurately and precisely than almost any human pilot, regardless of their experience or level of proficiency. That is NOT an indictment of human weakness, but rather an acknowledgment of automation benefits. Humans excel when “pre-programmed algorithms” come up short or they or the systems fail outright, such as creative problem-solving for unique issues, and abstract-thinking based decision-making are required. While fully automated aircraft control might be many years away, the building blocks of providing for it are mostly here today and being put into service as we write these posts. The ideal fully automated system will simultaneously appropriately leverage the strength of BOTH humans and machines while minimizing their respective weaknesses, without any implicit expectation of one to unknowingly “cover for” or otherwise “guess” what the other is doing during any portion of aircraft operation.

    To facilitate creation of this system, the professionals of aviation, the avionics designers, FAA legislators, pilots, and controllers, need to be apart of a on-going and evolving holistic and inclusive design process and decision making process, not one balkanized to each distinct group, whose resultant works are then legislatively foisted upon the other and told “deal with it, and learn it”, as that simply sets us all up for failure, and resultant problems, like the Asiana crash. The “accident chain” for that accident started many months, arguably years, earlier with Boeing’s automation design philosophy, and coupled with Korean culture and apparently insufficient recurrent training and oversight. Clearly there are problems, as I think we can all agree that that accident should NEVER have happened to a competent, properly trained, proficient crew, flying a fully-functional aircraft, to an unobstructed sea-level airport with an 11000 ft runway, on a cloudless, windless day.

    For now, the human PIC is still ultimately in charge, and ultimately responsible, and EACH needs to decide how they will ensure requisite systems and airmanship KNOWLEDGE and PROFICIENCY so that every takeoff results in a safe landing, regardless of the tools at their disposal, or decisions that have to be made promptly, such as taking controls from automation and/or human pilot, and “going around”. There may be valid reasons for egregious human-error accidents, but likely no valid excuses, and I’m pretty certain “I don’t know what the system is doing” is not one of them.

  • http://N/A Jim Wallace

    I flew for one of the regionals back in the 90’s, in the Beech 1900, (D model)
    and those aircraft were ordered from the factory with no autopilots. I saw it as a cost saving move at the time,…and some pilots complained about it, but you know, looking back, we all were made better pilots for having to hand fly every
    leg, and every approach! And when you flew your check ride, you had to show
    proficiency without the aid of the auto-pilot,….because there wasn’t one. : )

  • Bruce Landsberg

    To Avi’s point, we agree – more isn’t better – better is better. Also, if you design something for the top half of the bell shaped curve of human performance there will be some failures. Don’t ask me how to define that :-)

    There are too many qualifying conditions in my view for an optimally designed system in too many cases today.With Asiana however, this one may be different. The auto-throttle is not something that usually invites a lot of confusion and that there have been no AT accidents on the B777 after more than a decade in service leads me to think there are some other factors here. The NTSB report should be most enlightening.

  • Ron Rapp

    I agree, the NTSB report will be a fascinating read. I just wrote about this recently myself.

    It’s interesting to note that of the four pilots on the flight deck, three were trained by the military. The pilot flying was the only one who was an “ab initio” trainee of the airline itself. Since they don’t have much use for VFR flying, perhaps he never did much of it and simply did not develop any comfort with purely visual flying.

  • John R

    Lots of good ideas have appeared in the comments I’ve read so far but I always fall back to judgement. Judgement is DEVELOPED with experience and practice. Purposeful training with an experienced pilot or CFI (maybe I should reverse that and say experienced CFI). The old “Spot Landing” practice is a good exercise. I had a student years ago who had his own nice airplane with an autopilot. We had a hard time just flying the airplane and not using the magic. Don’t get me wrong, I really like the automation and use it a lot. Just remember it has no conscience.

  • Pranedh

    When close to the ground, switch off automation, scan the gauges and hand-fly. I believe the majority of pilots do that.

  • Glenn Swiatek

    A lot of good comments, by extremely knowledgeable pilots. I know after a long flight I’m not quite as good. But one concept not addressed here is very simple – competence.

    The freighter landing at the wrong Wichita airport had a better ending. In both cases the PIC was incompetent. In both cases there was another so called pilot sitting next to him, acting as a FAA std 170 lbs of dead weight.

    The good news is aviation’s safety record is so stellar many in the business are awaiting an NTSB report whose reccomendations will attempt to answer the age old question, how do you fix stupid ?

    By putting together another committee ? Great job security for the committee able.

  • unitechplctrainingseo

    Your blog are impressive to each other.I read your blog. useful content .its very good, Help ful for all. Thank you for sharing this.
    Industrial Automation Training Bangalore
    PLC & SCADA Training in Bangalore

  • scaindia

    There are a number of institutes in Bangalore that offer the PLC and SCADA training in Bangalore. Browse the web to avail info about the best PLC And SCADA training.

  • cymbio

    Your blog is very interesting. nice to read. Thanks for sharing
    Please visit the website:
    Green Coffee Bean Extract Manufacturers

  • sam sibells

    A couple of things. The bottom line on the Asiana crash is that the airspeed was not monitored. A basic requirement for flying even if on autopilot and/or using autothrottle.

    As for the PIC being stressed about doing a visual approach, I suspect that there may be more to this than we are allowing. He didn’t say why he was stressed as far as I know and that is an important question.

    One significant possibility is the reason why I might be stressed for doing a visual approach as he was. Not because of my worry about not being able to do it but because of where I work. In an environment that is likely similar to his. In a company with a punishment culture to enhance safety.

    Let me give you an example. Max landing weight in a particular widebody jet might require a speed that will give a 900 fpm descent rate. The 1000 fpm descent rate as advocated by the flight safety foundation is our maximum allowable. A pilot gets slightly high on approach, which is not uncommon and a bit of an increase in rate of descent is required. Nest thing you know, you were at 100 fpm for a few seconds.

    This is all recorded and sent back to the company. A few days later after your perfectly safe approach, you are now in trouble having to explain why you continued with an unstable approach below 1000 feet when in fact it was a quite safe approach. Then you get demoted in the name of safety. Yes this happens.

    You could have gone around of course as that is somehow safer although Oakland is you alternate so your extra fuel which you overburned on the way to destination is not much at all. And the weather is moving in. So of course it was safe to continue and land but the company sees things in black and white. An exceedence is an exceedence. I suspect Asian is like this.

    So what do you do to protect your livelihood, you say screw it, I am letting the automatics get me onto final. I know I could do my occasional opportunity to hand fly in a busy airspace but why. I suspect Asian is like this.

    And as for the Flight Safety Foundations well researched idea that anything more than 1000 fpm descent rate is unsafe, doesn’t make sense. This gave me a 500 fpm margin on the turboprops but sometime almost no margin for error on a big jet at max landing weight.

    Bottom line, I think his stress enroute about the upcoming approach may have come from an increased risk of encountering the punishment culture in the name of safety. Which was then followed by some significant errors.