Many airlines have policies against hand-flying the airplane. Like the discussion with self-driving cars; the computer is much safer overall to fly than a human pilot. However, the failure modes of the computer don't resemble the failure modes of the human pilot.
Since human pilots (on commercial flights) are either discouraged or prohibited from actually flying (for legitimate safety reasons), when the computer does experience a failure, the human is unskilled to take over. Yet, ironically, it appears that if we encouraged the human pilots to hand fly more, the accident rate would probably increase.
which is probably the FAA's point. Better to have an unskilled pilot not touching the commands than to have a skilled but aggressive pilot taking over and making mistakes with the lives of its passengers.
That's why the situation with computer-driven cars right now is so strange to me. Even if they're not free of flaws yet, they seem demonstrably much better drivers than humans already, because humans are so bad at it. But everything seems to be proceeding very slowly and cautiously, even though the number casualties from accidents warrants the opposite.
We are much more accepting of human errors. Machines have to have a reliability that is many orders of magnitude higher than humans. Just look at rail. We go to extreme lengths to make the rail system safe. Everything is redundant, every light has an extra wire that checks whether it's actually working. We could save a lot on railway infrastructure if we were ok with less than "perfect".
I see not only Google, but a number of auto manufacturers, doing a lot of work. The Google testing, which is on a relatively small number of precisely mapped roads in good weather, has had a quite a few incidents that required a human to intervene (which won't be a realistic option once normal drivers turn over control). So from my perspective there's quite a bit of work going on but a lot more experience and work is needed before this is something that can be sold to consumers even for limited use cases.
(What we are already getting in high-end vehicles is assistive driving systems which should reduce accidents as they become more widespread--at least so long as they're not so effective as to increase issues associated with inattentiveness.)
Imagine if self-driving cars were at a point where they were 10% better (in terms of fatalities caused) than human drivers. You start flooding the roads with them, and fatalities decrease.
Then the news stories start. "Mother of Four Killed by Self-Driving Car." Video of the grieving children. News shows host "experts" who go on at length about the deficits in computer judgment. The family sues, with their lawyer saying that this death would not have happened with a human at the controls. (And he might be right! Even if better than humans, computers will probably get into different accidents, not just fewer.) Restrictive regulations or even bans start being discussed in Congress.
Maybe it wouldn't happen this way, but it could potentially set things back by years. Looking at the technology, pushing for early adoption might save lives. But looking at the politics, pushing too hard could cost lives, by making people unwilling to accept the technology.
There's a lot required before they're ready in any case. (How much probably depends a fair bit on where and under what circumstances they're allowed to be on full auto-pilot.) But you're absolutely right on your basic point. And even if they're safer than humans overall, if they have a tendency (even if rarely) to get into serious accidents in which they self-evidently did something dumb (from a human's perspective), that would be a big issue.
It's interesting to imagine an extreme case. Imagine if self-driving cars were absolutely perfect drivers, except every 100 billion miles the car accelerates to top speed and then drives straight into a bridge support.
Such a car would be immensely safer than what we have now. Would people willingly ride in one?
I think I'm pretty good at rationally weighing the risks, but just imagining this still gives me the willies.
> Maybe it wouldn't happen this way, but it could potentially set things back by years. Looking at the technology, pushing for early adoption might save lives. But looking at the politics, pushing too hard could cost lives, by making people unwilling to accept the technology.
It would absolutely happen this way.
Like any change, driverless cars would cause a lot of people to lose money and/or their livelihoods. So there will be a significant number of people (and businesses) who fight against the change.
Look at climate change -- the actual metrics don't even matter. There is always a way to skew the display of the numbers to make a pretty graph that will convince enough of the population that there's at LEAST some legitimate controversy over the issue; which is enough to just put the decision off until the controversy is over.
I've heard something similar with safety products on cars. Alert systems that warn you about things in your blind spot or potentially rear-ending someone causes people to be less aware/more relaxed than they should be.
> Yet, ironically, it appears that if we encouraged the human pilots to hand fly more, the accident rate would probably increase.
An interesting analog is the history of spin training in the US. Until the 1950s, a pilot was required to enter and recover from a 3 turn spin on their private pilot checkride. The thought was spins are dangerous, so pilots should show they can recover. The problem is that spins ARE dangerous, so a large number of students and instructors were spinning into the ground during training.
The FAA decided to remove the spin from the checkride. This effectively removed it from training. Many instructors lamented this, claiming the spin accident rate among private pilots would increase dramatically. They couldn't be expected to recover from something they had never trained for. The accident rate did change. It dropped sharply. But why?
The first part is the FAA accomplished its goal. Less pilots intentionally spinning meant less pilots spinning into the ground. But this didn't explain the drop in spin related accidents in nontraining situations.
It all came down to the airplanes. The airplane that most pilots train in is the same kind of airplane that most general aviation pilots fly: small, slow, single engine two or four seaters. Airplane manufacturers designed airplanes that could fulfill the needs of both the student pilot and weekend pilot. This meant a school could rent out its training airplanes, and a pilot could learn in the airplane he meant to own long term.
One of the requirements of the student's airplane was that it could be used for spin training. This meant making an airplane easy to spin. The problem here is clear in retrospect. You have a population of weekend pilots flying airplanes that are specifically designed to spin easily!
Removing the training requirement removed the manufacturing requirement. There is a clear difference in the design of airplanes before and after this change. Compare the Cessna 140 and 150, or the Piper Super Cub and Piper Cherokee. The later model is incredibly more difficult to spin than the earlier one.
Spin accidents in these post change airplanes are almost nonexistent. The FAA statistics are even misleadingly negative. They lump stall/spin accidents together. But, if you look into the accident reports in detail, you'll find a fully developed gyroscopic spin into the ground is incredibly rare.
This story can serve as an example for other interactions between regulations, operators and machine. Remove training requirements that do more harm than good. Focus on emergency avoidance instead of emergency recovery. Allow engineers to create the safest possible design, even one that doesn't pass archaic requirements.
No pilot, ship captain, or driver operates in a vacuum. If we want the best possible environment, we have to look at the environment as a whole.
Hopefully, as simulators get closer and closer to real life (and become cheaper), pilot in training will be able to train for those events without actually risking them in real life.
Removing the ability/encouragement to hand fly the airplane in real life - as some Asian/ME carriers have done either through punitive QAR (Quick Access Recorder, like the Flight Data Recorder) and FOQA (Flight Operations Quality Assurance aka spy on the pilot) regimes that track when a pilot blows an approach or other parameter - will only make pilots MORE dependent on automation and unable to cope when automation fails, ala Air France 447. [1]
That's also partially why we had Asiana 214 - if the pilot at the controls was more skilled at hand flying the aircraft he would not have been so dependent on the autothrottle functionality to manage the speed of the airplane.
I worry deeply about creeping automation not only on the flight deck but in our cars. Although automation is far more responsive and adds greatly to safety in both cases, dependence on automation will breed a new generation of the "Children of the Magenta" [2] who blindly allow the computer to follow the magenta line on the display, either into the ground in an airplane, or into oncoming traffic in a car as they ask themselves "what's it doing now??!"
There's some interesting academic research going on about the interactions between humans and computer decision making. One place I know of is the Humans and Autonomy lab at Duke http://hal.pratt.duke.edu/
I saw an interesting panel last year at the MIT Sloan CIO Symposium where Missy Cummings from this lab (who used to by a navy fighter pilot) talked about a number of these issues.
In many ways the weekend warrior that flies his Cessna 172 around for fun is probably a better pilot that a lot of guys flying the big jets. One hand flies everything the other mostly just kicks on the autopilot and sits back.
There are certainly airline pilots that still have a passion for aviation... some even still teach in their down time, but it's increasingly uncommon. Too many don't do enough 'real' flying anymore. They're mostly just sitting up front to make sure the autopilot doesn't screw up... and sadly accident data shows when the autopilot does screw up the pilot forgets what to do (see accident reports where the autopilot stalls the plane and pilots forget basic stall recovery procedures).
What the heck did "Emotionally distraught family members grasped at resolution while the pilots of the crash on the Hudson supported solutions founded in myth" mean? Perhaps this author should spend less time flying planes and more time learning how to write.
Making "the crash on the Hudson" a link would've helped a lot. I guess for that the article's main audience, aviators and aviation bureaucrats, this is well known?
I know about the landing on the Hudson (my apartment faces the Hudson and I saw it happen), but this paragraph still didn't make sense to me. What solution did the pilots push? Sullenberger, the pilot on that flight is both a glider pilot and a GA flight instructor, so he seems to be the perfect argument that GA skills lend themselves to better problem resolution in unexpected flight situations: he was essentially landing a glider when both engines went.
I was lucky enough to have a father as a pilot who paid for me to get my private pilots license. Despite the exorbitant price tag, I look back on that education as immensely and uniquely beneficial to my development.
As per his demands, I would complete a solo circuit* in an airplane before I drove a car solo. It didn't help that the airport I was learning was nestled in a tight mountain valley and notorious for it's shifty winds. The best lesson I learned was one of immense personal responsibility. It didn't matter what caused the accident if you ended up dead. That kind of responsibility-at-all-costs mentality has stuck with me to this day. Don't put yourself in a position where something or someone else can cause you to suffer. Check that the fuel attendant put the right amount of fuel you asked for in your airplane. Check that the mechanic wired the flight controls correctly. Check that the weather forecast is rational. The meteorologist isn't going to climb into your airplane and fly into that region-- you are.
This holds true for every part of my life. I don't follow cars close enough so that I depend on their judgement.
If you've got the money, I strongly recommend anyone to get their wings. It's the best education I've ever received.
* American pilots will know this as a 'pattern': takeoff, loop around the airport and immediately land. The shortest complete flight possible.
Your diligence was not unwarranted. Before anybody runs out to get their wings, realize that general aviation is quite dangerous. Here are two links that use similar math to arrive at the same 20x-more-dangerous-than-driving figure:
"flying is no more than 10 times as dangerous per mile of travel."
That may sound like flying is considerably more dangerous than driving, but this is an overall statistic. Most GA accidents happen because the pilot did something stupid like taking off without enough fuel, or flying into weather they weren't equipped to handle. So if you avoid doing stupid things, you can stack the odds considerably in your favor. The risk of flying can be comparable to the risk of driving.
Driving is dangerous, and if it takes "stacking the odds considerably in your favor" to make $THING only as dangerous as driving, I think it's safe bet that $THING is a pretty dangerous activity.
I'm actively working toward my PPL now but let's not pretend it's a safe venture by any means. You're strapping yourself onto a single lawn-mower engine that was build in the 60's that someone bought for $20,000 five years ago.
Of course, the same applies to driving. Take out DUIs, excessive speeding, texting, etc. and the accident rate associated with driving goes way down as well.
The problem is that when you're driving your fate is much more in the hands of the drivers around you than when you're flying. A stupid driver often takes out nearby cars along with their own. It's extremely rare for a stupid pilot to take out any aircraft but his own.
While true, you can still heavily tilt the odds in your favor while driving by being attentive, being in control, and avoiding bad weather when possible.
And, on the flip side, it's unlikely that another plane is going to run into you but you're probably a lot more exposed to events out of your control such as unexpected weather (e.g. crosswinds) and mechanical failures with serious consequences in a plane than in a car.
That's sort of the debate here, I think. Given $DANGEROUS_THING with $EVENTS_BEYOND_MY_CONTROL, I want to predict and plan for $EVENTS_BEYOND_MY_CONTROL to the best of my ability.
What's more unpredictable: stupid human drivers, stupid robot drivers, complex mechanical machines, or weather? Who is John Galt?
Hmm, I tell myself the same things about my driving ("I avoid doing stupid things, so I have stacked the odds in my favor.") -- possible since the causes of most car accidents are well-known.
I guess in both cases, we believe we have improved the odds, but also in both, we don't know how much.
There is the additional problem that our efforts to improve our own safety may be less airtight than we imagine.
That's true, but not particularly actionable. "Avoid doing stupid things" isn't something you can just decide to do directly. It's a good goal, but you need to find more specific ways to achieve that goal.
There are long established ways to reduce the incidence of stupidity. Always using checklists, always getting a weather brief, establishing personal minimums, etc. Of course, all of this will be covered as a standard part of any flying safety course.
IMO just saying "avoid doing stupid things" is dangerous. It's too easy to take that advice and say, well, I'm obviously not stupid, so I'm safe. And then you fly into a thunderstorm or stall while turning final or forget to connect a control surface, because avoiding stupidity has little if anything to do with being smart.
You can definitely stack the odds heavily in your favor, but it takes a lot of effort.
general aviation is only dangerous because of person between the controls and seat. Modern engines don't fail and the likelihood of exogenous events occurring are very very remote. Should you properly contain errors caused by pilot error, flying is remarkably safe.
Can't we build our own propensity for imperfect decision making into a calculation of safety? Can I not realize that while I consider myself more rational and cautious than my neighbors, they likely feel the same, and X% of them die when they try to fly?
I would consider this "second derivative caution". It got cemented into me when I was taking hang gliding lessons and saw one of the most stone cold cautious members of my class become exuberant and act rashly on a landing, and break his arm quite badly.
That's not quite true. Outright structural failure is not common, but there is a small but non-negligible category of accidents that are originate in improper maintenance, contaminated fuel, and similar non-pilot related causes. And piston aircraft engines have a disturbing tendency to fail compared to turbine engines or even automobile engines.
If you draw the line that it's the pilot's responsibility to ensure that the aircraft is airworthy then I guess you can place the proximate blame of that on the pilot, too. By that rationale, however, all car accidents are also your fault, because of the catch-all rule that says you may not drive so fast so as to put "persons or property at risk" or however it's phrased.
improper maintenance, contaminated fuel, and similar non-pilot related causes.
Before you ever takeoff, you must check for these things.
> maintenance:
do a full walk-around of the aircraft, checking for any damage. You can't catch everything, but you should be thorough enough so that if anything sudden did happen, ie: your airplane got clipped by the fuel truck, that you would catch it.
> contaminated fuel
Aviation fuel is dyed blue. Before every flight we take a small sample of fuel from the bottom of each tank to ensure water hasn't contaminated our fuel. On highwing aircraft we also draw fuel from the fuel lines leading down into the engine. This is sufficient to almost completely rule out fuel contamination.
I'm not saying that sudden accidents don't occur, merely that they are extremely remote. Even then, it is the responsibility of the pilot to plan accordingly and to further mitigate the consequences of mechanical breakdowns.
The production system for building most cars, I'm not sure about planes, is heavily dependent on built in quality. There are final quality checks, but not a comprehensive double checking of everything. The assumption is someone or some process will not pass on a defect. This is different from pre-flight checks though.
The enforcement focus for the FAA seems to be changing, though. The new FAA administrator has stated that they will move from the punishment-focused enforcement to a "standards compliance" focus, where education and training can be a more effective tool than just enforcement. Seems like a promising change.
What the author specifies in the article all makes sense. There's one important thing missing, however- any evidence to support his argument.
I didn't see a single data point that indicates that commercial aircraft are less safe than they were at any previous point in time. With more flights taking off now than 40 years ago and avoidable crashes occurring due to pilot error being few and far between, this seems like nothing more than a guy with an opinion pontificating on a blog as he gazes out his window.
Right; and the biggest high-profile accidents in the last decade have been caused when the complexity of the automation overwhelmed the pilots and caused them to make bad decisions because they did not know if they could trust the values the computer was giving them.
Basic flying skills won't help you fix that -- if you're over the middle of the ocean at night at altitude in the middle of a storm with strong updrafts and your instruments are giving you strange readings, it can be really hard to "feel" whether you're actually in a stall. Especially if your airspeed indicators had been acting up earlier in the flight, so you knew you couldn't trust that reading. It's ultimately an interface design problem: with new control systems, it's not always apparent what interface will be required in every edge case.
But ultimately, you're right. Flying is so safe now that major commercial airline accidents require 5 or 6 cascading failures in many different parts of the system (maintenance, design, security, pilots, ATC, etc) to occur. Even in the accidents I mention above there were other failures that happened outside the plane: the plane wasn't de-iced properly, or ATC routed them through a storm they really shouldn't have.
I have a huge fear of flying (that seems to be increasing), and it's beginning to affect my life. For example, I have the opportunity to go to Portland for a conference at the end of the month but I am convincing myself not to go because I don't know if I can handle a two 6 hour flights in a span of 2 days. I literally have another tab open on delta.com with flights that I should book, but it's more likely that I won't.
I have a doctor who has prescribed me Xanax strictly for flying, in which case I take a lot and it sort of just knocks me out. The problem however is that roughly three weeks out, if I know I have a flight, I will start to really stress, obsess about the upcoming flight, have difficulty sleeping, etc... And like most of us here on HN, I consider myself a logical person who understands the safety of flying, the statistics and the likelihood of events, yadda yadda, but it doesn't change the fact that for me, I feel extremely uncomfortable flying. I don't know whether it's the lack of control or the anonymity of whoever is in the cockpit, but sitting on a commercial jet coincides with me feeling panic.
Yeah I recently read When Panic Attacks by David Burns, which is all about CBT, and while I see the value in what he is saying it really didn't change my mindset about things. There is a separation between what is rational and what is mental/emotional. I actually feel like I could do a pretty good job talking to someone else who has a fear of flying and helping them with what my logical side knows, but internally I am still a wreck.
Interesting - I will definitely look into that. I remember having no fear of flying the first time I flew (when I was like 7 or 8), but I remember looking at my mom during takeoff, white-knuckling the armrest. I've also foolishly done a lot of research over the years on airline disasters, but I think from there on out something in my mind switched from this is fun and interesting to this does not feel safe.
One of the things you get from ACT is; if your mind can switch from "this is fun and interesting" to "this does not feel safe" then it can just as easily switch back.
Yeah, I got a fear of flying also. I live and work in one end of Europe, but come from another end, so when I go back home for the holidays I ride a couple of trains and take a long boat trip at the end. It takes me three days to get home, but I don't have to think about flying. Think about it - it could work for you too.
To be honest, lately I started flying again after almost ten years I avoided it like the plague. It's not that I'm less afraid of it, but it's eating away so much of my holiday that I just can't afford not to anymore.
I guess in a way you could say that getting sick of the long journey by land and sea did help me get over my fear of flying a little bit.
Fear of flying is a strange thing. I never had any issues until I happened to be on a regional flight on one of those tiny little Embraer three-seat-wide jets. We are probably 10 minutes away from starting descent, when suddenly I have the first real panic attack I have ever experienced: It feels as if some part of my lizard brain has started spamming the “IMMINENT MORTAL DANGER GET OUT NOW” signal, throwing the rest of my brain into chaos trying to form a coherent threat model. It seems to have something to do with the claustrophobic cabin and the lack of direct control, but everything becomes a fever-dream blur.
I don’t really remember how the episode ended other than that I was ultimately able to walk off the plane unaided, but now whenever I fly, I triple check I’m not on a compact airliner like that.
I would really recommend seeing a therapist that practices cognitive behavioral therapy. I was where you are. Reading about CBT didn't help. I had to go and talk to a professional. It hasn't cured me but I can consistently fly at a much lower level of anxiety and the anticipatory anxiety is down to the day of.
> A decade later, many pilots struggle to perform the most basic and critical maneuver, a stall recovery, and ...
I just flat out don't believe that modern airline pilots have more trouble with stall recovery than pilots of earlier years.
Now I might be wrong, I'm just a layperson here, but my understanding is that airline pilots have to practice stall recovery in flight simulators. A flight simulator can't be 100% accurate to a real aircraft's behavior in a real stall, but my impression is that modern flight simulators are quite realistic.
Most of the emergency training that jet pilots go through are system failures -- engine failures, fires, and whatnot. Also, a stall in a sim is a very different experience than a stall in a real plane because in the back of your mind you know that your life is not on the line in a sim. That emotional factor can make a big difference when you have to recover from an unintended stall in a real airplane.
I think the problem, in cases like the AF crash into the Atlantic, is that modern planes are designed to not be possible to fly into a stall. So when some sensor fails and it suddenly becomes possible, the pilots don't even realize they are stalling the plane.
These threads come up often in various news aggregation sites, and I feel that the comments tend to overflow with people that have no commercial flying experience stating things like, "commercial pilots don't do REAL flying like small aircraft pilots do, they just turn on the autopilot and watch."
This is only true in the sense that there is less hands-on-the-flight-controls time. Flying a large commercial aircraft is actually very difficult - so difficult, in fact, that the menial tasks like physically handling the controls can be better delegated to an autopilot most of the time. The control of the aircraft has been abstracted out into much higher-level management of systems. It's more like the manager who has one eye on where the company is headed and one eye on what tasks his employees are completing, and it is his responsibility to make sure those things match.
I try to explain this relationship to my friends by likening the autopilot to something more like cruise control in the car: when you turn on cruise control, it allows you to take your foot off the gas pedal, but by no means does it mean you aren't actively managing the safe outcome of the trip; it's just a different tool. Sure, autopilots are more complex than cruise control, but that doesn't adulterate my main point.
Don't get me wrong, it has always been a struggle to make sure that pilots of large aircraft get enough stick time and emergency training; you have to fit recurrent training into an already very busy schedule and spend lots of money on expensive simulators. That being said, the FAA (USA) regulates pretty heavily how much and exactly what types of recurrent training is necessary to maintain your position, and stall recovery is one of them.
This statement: "many pilots struggle to perform the most basic and critical maneuver, a stall recovery" seems like pure applesauce, and if they want to back that up with some better data about how that has changed over time and how it has affected airline safety over the years, I'd love to see it. We are in a relatively safe period of airline travel, and I believe we will continue to be.. http://www.cnn.com/interactive/2014/07/travel/aviation-data/
Note that the author is an "airport owner". So he has a direct interest in more general aviation.
Airline pilots today rarely fly light aircraft recreationally because they can't afford to. Airline pilots are paid much less than they used to be.[1] Regional airline pilots make $30K - $60K. Some bus drivers are paid more. Only senior pilots of major airlines break $100K today. Also, in the US, pilots usually have to pay for their own training.
As for "losing the basic skills", the author doesn't mention the best known example - the crash of Asiana Airlines Flight 214 at SFO. The pilots had to make a manual landing in clear weather, instead of using autoland, and they undershot the runway.
Some of that was a lack of manual piloting skills, and some of that was mode trouble with cockpit automation. (This is a huge issue. Read crash reports, and all too often part of the problem was something being in the wrong mode.) On the other hand, the "miracle on the Hudson" of US Airways Flight 1549, which he does mention, was a triumph of automatic control - once the pilot decided to land on the river, the Airbus control system did a very good job of providing a smooth descent.[2]
> Also, in the US, pilots usually have to pay for their own training.
This is half true. If you're in HS and you want to get a job as a commercial pilot (ATP) right out of the gate, yes you will need to pay for your flight training.
However, the vast majority of ATPs received their flight training in the military, for which they obviously did not have to pay. Part of this is the sheer number of hours you need to even apply for ATP jobs.
It should be noted that Capt. Sullenberger was one of those apparently increasingly rare birds, an airline captain with extensive and current hands-on flying experience.
I was not aware of Langewiesche's book until now, so perhaps you can tell us if the automation handled the flare, which I believe would have been the most critical stage of this landing (if the fuselage had broken, as has happened in many other ditchings, it is unlikely that everyone would have made it out.) If so, did those systems compensate for the fact that the wheels were not down and so touchdown would be made on the engine nacelles with the fuselage closer to the surface than for a normal landing? From the video, it appeared to be a near-perfect three-point landing, which is not the attitude an airliner normally lands in.
Following on from my question about Langewiesche's analysis, my understanding is that aside from losing propulsion, the aircraft was largely undamaged. Given decent visibility and somewhere to put it down, an unpowered landing of an airliner would be considered challenging but doable with a high probability of success, even without control-system automation - see the Gimli Glider.
>However, what if you were a passenger on a commercial jetliner that had run out of gas?
Probably a reference to the Gimli Glider. Airliner ran out of fuel and the pilot used his general aviation background to forward slip it on to an auto race track.
Myabe i was just paranoid, but i quit flying private after about 1000 hours because of the ever present opportunity to make a mistake and have the FAA come down on me. A private pilot is under the bloodshot eye of the FAA from the time he sets foot on the airport to when he leaves it.
69 comments
[ 3.1 ms ] story [ 156 ms ] threadSince human pilots (on commercial flights) are either discouraged or prohibited from actually flying (for legitimate safety reasons), when the computer does experience a failure, the human is unskilled to take over. Yet, ironically, it appears that if we encouraged the human pilots to hand fly more, the accident rate would probably increase.
I see not only Google, but a number of auto manufacturers, doing a lot of work. The Google testing, which is on a relatively small number of precisely mapped roads in good weather, has had a quite a few incidents that required a human to intervene (which won't be a realistic option once normal drivers turn over control). So from my perspective there's quite a bit of work going on but a lot more experience and work is needed before this is something that can be sold to consumers even for limited use cases.
(What we are already getting in high-end vehicles is assistive driving systems which should reduce accidents as they become more widespread--at least so long as they're not so effective as to increase issues associated with inattentiveness.)
Imagine if self-driving cars were at a point where they were 10% better (in terms of fatalities caused) than human drivers. You start flooding the roads with them, and fatalities decrease.
Then the news stories start. "Mother of Four Killed by Self-Driving Car." Video of the grieving children. News shows host "experts" who go on at length about the deficits in computer judgment. The family sues, with their lawyer saying that this death would not have happened with a human at the controls. (And he might be right! Even if better than humans, computers will probably get into different accidents, not just fewer.) Restrictive regulations or even bans start being discussed in Congress.
Maybe it wouldn't happen this way, but it could potentially set things back by years. Looking at the technology, pushing for early adoption might save lives. But looking at the politics, pushing too hard could cost lives, by making people unwilling to accept the technology.
Such a car would be immensely safer than what we have now. Would people willingly ride in one?
I think I'm pretty good at rationally weighing the risks, but just imagining this still gives me the willies.
It would absolutely happen this way.
Like any change, driverless cars would cause a lot of people to lose money and/or their livelihoods. So there will be a significant number of people (and businesses) who fight against the change.
Look at climate change -- the actual metrics don't even matter. There is always a way to skew the display of the numbers to make a pretty graph that will convince enough of the population that there's at LEAST some legitimate controversy over the issue; which is enough to just put the decision off until the controversy is over.
An interesting analog is the history of spin training in the US. Until the 1950s, a pilot was required to enter and recover from a 3 turn spin on their private pilot checkride. The thought was spins are dangerous, so pilots should show they can recover. The problem is that spins ARE dangerous, so a large number of students and instructors were spinning into the ground during training.
The FAA decided to remove the spin from the checkride. This effectively removed it from training. Many instructors lamented this, claiming the spin accident rate among private pilots would increase dramatically. They couldn't be expected to recover from something they had never trained for. The accident rate did change. It dropped sharply. But why?
The first part is the FAA accomplished its goal. Less pilots intentionally spinning meant less pilots spinning into the ground. But this didn't explain the drop in spin related accidents in nontraining situations.
It all came down to the airplanes. The airplane that most pilots train in is the same kind of airplane that most general aviation pilots fly: small, slow, single engine two or four seaters. Airplane manufacturers designed airplanes that could fulfill the needs of both the student pilot and weekend pilot. This meant a school could rent out its training airplanes, and a pilot could learn in the airplane he meant to own long term.
One of the requirements of the student's airplane was that it could be used for spin training. This meant making an airplane easy to spin. The problem here is clear in retrospect. You have a population of weekend pilots flying airplanes that are specifically designed to spin easily!
Removing the training requirement removed the manufacturing requirement. There is a clear difference in the design of airplanes before and after this change. Compare the Cessna 140 and 150, or the Piper Super Cub and Piper Cherokee. The later model is incredibly more difficult to spin than the earlier one.
Spin accidents in these post change airplanes are almost nonexistent. The FAA statistics are even misleadingly negative. They lump stall/spin accidents together. But, if you look into the accident reports in detail, you'll find a fully developed gyroscopic spin into the ground is incredibly rare.
This story can serve as an example for other interactions between regulations, operators and machine. Remove training requirements that do more harm than good. Focus on emergency avoidance instead of emergency recovery. Allow engineers to create the safest possible design, even one that doesn't pass archaic requirements.
No pilot, ship captain, or driver operates in a vacuum. If we want the best possible environment, we have to look at the environment as a whole.
That's also partially why we had Asiana 214 - if the pilot at the controls was more skilled at hand flying the aircraft he would not have been so dependent on the autothrottle functionality to manage the speed of the airplane.
I worry deeply about creeping automation not only on the flight deck but in our cars. Although automation is far more responsive and adds greatly to safety in both cases, dependence on automation will breed a new generation of the "Children of the Magenta" [2] who blindly allow the computer to follow the magenta line on the display, either into the ground in an airplane, or into oncoming traffic in a car as they ask themselves "what's it doing now??!"
[1] https://en.wikipedia.org/wiki/Air_France_Flight_447
[2] https://www.youtube.com/watch?v=pN41LvuSz10
I saw an interesting panel last year at the MIT Sloan CIO Symposium where Missy Cummings from this lab (who used to by a navy fighter pilot) talked about a number of these issues.
There are certainly airline pilots that still have a passion for aviation... some even still teach in their down time, but it's increasingly uncommon. Too many don't do enough 'real' flying anymore. They're mostly just sitting up front to make sure the autopilot doesn't screw up... and sadly accident data shows when the autopilot does screw up the pilot forgets what to do (see accident reports where the autopilot stalls the plane and pilots forget basic stall recovery procedures).
As per his demands, I would complete a solo circuit* in an airplane before I drove a car solo. It didn't help that the airport I was learning was nestled in a tight mountain valley and notorious for it's shifty winds. The best lesson I learned was one of immense personal responsibility. It didn't matter what caused the accident if you ended up dead. That kind of responsibility-at-all-costs mentality has stuck with me to this day. Don't put yourself in a position where something or someone else can cause you to suffer. Check that the fuel attendant put the right amount of fuel you asked for in your airplane. Check that the mechanic wired the flight controls correctly. Check that the weather forecast is rational. The meteorologist isn't going to climb into your airplane and fly into that region-- you are.
This holds true for every part of my life. I don't follow cars close enough so that I depend on their judgement.
If you've got the money, I strongly recommend anyone to get their wings. It's the best education I've ever received.
* American pilots will know this as a 'pattern': takeoff, loop around the airport and immediately land. The shortest complete flight possible.
http://aviation.stackexchange.com/questions/198/how-dangerou...
http://philip.greenspun.com/flying/safety
That may sound like flying is considerably more dangerous than driving, but this is an overall statistic. Most GA accidents happen because the pilot did something stupid like taking off without enough fuel, or flying into weather they weren't equipped to handle. So if you avoid doing stupid things, you can stack the odds considerably in your favor. The risk of flying can be comparable to the risk of driving.
I'm actively working toward my PPL now but let's not pretend it's a safe venture by any means. You're strapping yourself onto a single lawn-mower engine that was build in the 60's that someone bought for $20,000 five years ago.
And, on the flip side, it's unlikely that another plane is going to run into you but you're probably a lot more exposed to events out of your control such as unexpected weather (e.g. crosswinds) and mechanical failures with serious consequences in a plane than in a car.
What's more unpredictable: stupid human drivers, stupid robot drivers, complex mechanical machines, or weather? Who is John Galt?
I guess in both cases, we believe we have improved the odds, but also in both, we don't know how much.
There is the additional problem that our efforts to improve our own safety may be less airtight than we imagine.
There are long established ways to reduce the incidence of stupidity. Always using checklists, always getting a weather brief, establishing personal minimums, etc. Of course, all of this will be covered as a standard part of any flying safety course.
IMO just saying "avoid doing stupid things" is dangerous. It's too easy to take that advice and say, well, I'm obviously not stupid, so I'm safe. And then you fly into a thunderstorm or stall while turning final or forget to connect a control surface, because avoiding stupidity has little if anything to do with being smart.
You can definitely stack the odds heavily in your favor, but it takes a lot of effort.
I would consider this "second derivative caution". It got cemented into me when I was taking hang gliding lessons and saw one of the most stone cold cautious members of my class become exuberant and act rashly on a landing, and break his arm quite badly.
If you draw the line that it's the pilot's responsibility to ensure that the aircraft is airworthy then I guess you can place the proximate blame of that on the pilot, too. By that rationale, however, all car accidents are also your fault, because of the catch-all rule that says you may not drive so fast so as to put "persons or property at risk" or however it's phrased.
Before you ever takeoff, you must check for these things.
> maintenance:
do a full walk-around of the aircraft, checking for any damage. You can't catch everything, but you should be thorough enough so that if anything sudden did happen, ie: your airplane got clipped by the fuel truck, that you would catch it.
> contaminated fuel
Aviation fuel is dyed blue. Before every flight we take a small sample of fuel from the bottom of each tank to ensure water hasn't contaminated our fuel. On highwing aircraft we also draw fuel from the fuel lines leading down into the engine. This is sufficient to almost completely rule out fuel contamination.
I'm not saying that sudden accidents don't occur, merely that they are extremely remote. Even then, it is the responsibility of the pilot to plan accordingly and to further mitigate the consequences of mechanical breakdowns.
See e.g. http://www.ainonline.com/aviation-news/business-aviation/201...
I didn't see a single data point that indicates that commercial aircraft are less safe than they were at any previous point in time. With more flights taking off now than 40 years ago and avoidable crashes occurring due to pilot error being few and far between, this seems like nothing more than a guy with an opinion pontificating on a blog as he gazes out his window.
Basic flying skills won't help you fix that -- if you're over the middle of the ocean at night at altitude in the middle of a storm with strong updrafts and your instruments are giving you strange readings, it can be really hard to "feel" whether you're actually in a stall. Especially if your airspeed indicators had been acting up earlier in the flight, so you knew you couldn't trust that reading. It's ultimately an interface design problem: with new control systems, it's not always apparent what interface will be required in every edge case.
But ultimately, you're right. Flying is so safe now that major commercial airline accidents require 5 or 6 cascading failures in many different parts of the system (maintenance, design, security, pilots, ATC, etc) to occur. Even in the accidents I mention above there were other failures that happened outside the plane: the plane wasn't de-iced properly, or ATC routed them through a storm they really shouldn't have.
I have a doctor who has prescribed me Xanax strictly for flying, in which case I take a lot and it sort of just knocks me out. The problem however is that roughly three weeks out, if I know I have a flight, I will start to really stress, obsess about the upcoming flight, have difficulty sleeping, etc... And like most of us here on HN, I consider myself a logical person who understands the safety of flying, the statistics and the likelihood of events, yadda yadda, but it doesn't change the fact that for me, I feel extremely uncomfortable flying. I don't know whether it's the lack of control or the anonymity of whoever is in the cockpit, but sitting on a commercial jet coincides with me feeling panic.
To be honest, lately I started flying again after almost ten years I avoided it like the plague. It's not that I'm less afraid of it, but it's eating away so much of my holiday that I just can't afford not to anymore.
I guess in a way you could say that getting sick of the long journey by land and sea did help me get over my fear of flying a little bit.
I don’t really remember how the episode ended other than that I was ultimately able to walk off the plane unaided, but now whenever I fly, I triple check I’m not on a compact airliner like that.
I just flat out don't believe that modern airline pilots have more trouble with stall recovery than pilots of earlier years.
Now I might be wrong, I'm just a layperson here, but my understanding is that airline pilots have to practice stall recovery in flight simulators. A flight simulator can't be 100% accurate to a real aircraft's behavior in a real stall, but my impression is that modern flight simulators are quite realistic.
This is only true in the sense that there is less hands-on-the-flight-controls time. Flying a large commercial aircraft is actually very difficult - so difficult, in fact, that the menial tasks like physically handling the controls can be better delegated to an autopilot most of the time. The control of the aircraft has been abstracted out into much higher-level management of systems. It's more like the manager who has one eye on where the company is headed and one eye on what tasks his employees are completing, and it is his responsibility to make sure those things match.
I try to explain this relationship to my friends by likening the autopilot to something more like cruise control in the car: when you turn on cruise control, it allows you to take your foot off the gas pedal, but by no means does it mean you aren't actively managing the safe outcome of the trip; it's just a different tool. Sure, autopilots are more complex than cruise control, but that doesn't adulterate my main point.
Don't get me wrong, it has always been a struggle to make sure that pilots of large aircraft get enough stick time and emergency training; you have to fit recurrent training into an already very busy schedule and spend lots of money on expensive simulators. That being said, the FAA (USA) regulates pretty heavily how much and exactly what types of recurrent training is necessary to maintain your position, and stall recovery is one of them.
This statement: "many pilots struggle to perform the most basic and critical maneuver, a stall recovery" seems like pure applesauce, and if they want to back that up with some better data about how that has changed over time and how it has affected airline safety over the years, I'd love to see it. We are in a relatively safe period of airline travel, and I believe we will continue to be.. http://www.cnn.com/interactive/2014/07/travel/aviation-data/
Airline pilots today rarely fly light aircraft recreationally because they can't afford to. Airline pilots are paid much less than they used to be.[1] Regional airline pilots make $30K - $60K. Some bus drivers are paid more. Only senior pilots of major airlines break $100K today. Also, in the US, pilots usually have to pay for their own training.
As for "losing the basic skills", the author doesn't mention the best known example - the crash of Asiana Airlines Flight 214 at SFO. The pilots had to make a manual landing in clear weather, instead of using autoland, and they undershot the runway. Some of that was a lack of manual piloting skills, and some of that was mode trouble with cockpit automation. (This is a huge issue. Read crash reports, and all too often part of the problem was something being in the wrong mode.) On the other hand, the "miracle on the Hudson" of US Airways Flight 1549, which he does mention, was a triumph of automatic control - once the pilot decided to land on the river, the Airbus control system did a very good job of providing a smooth descent.[2]
[1] http://thetruthabouttheprofession.weebly.com/professional-pi... [2] http://www.amazon.com/Fly-Wire-Geese-Miracle-Hudson/dp/03126...
This is half true. If you're in HS and you want to get a job as a commercial pilot (ATP) right out of the gate, yes you will need to pay for your flight training.
However, the vast majority of ATPs received their flight training in the military, for which they obviously did not have to pay. Part of this is the sheer number of hours you need to even apply for ATP jobs.
I was not aware of Langewiesche's book until now, so perhaps you can tell us if the automation handled the flare, which I believe would have been the most critical stage of this landing (if the fuselage had broken, as has happened in many other ditchings, it is unlikely that everyone would have made it out.) If so, did those systems compensate for the fact that the wheels were not down and so touchdown would be made on the engine nacelles with the fuselage closer to the surface than for a normal landing? From the video, it appeared to be a near-perfect three-point landing, which is not the attitude an airliner normally lands in.
Probably a reference to the Gimli Glider. Airliner ran out of fuel and the pilot used his general aviation background to forward slip it on to an auto race track.
* https://en.wikipedia.org/wiki/Gimli_Glider
Awesome instructors. Just got a job at LAN peru.