"The Air France story is a story about a failed handoff, where the automation onboard an airplane found a relatively minor fault and handed control of the plane back to the human pilots, too suddenly and ungracefully surprised them. And they had lost some of their skills flying too much with the automated systems and lost control of the airplane. Which actually was a perfectly good airplane about a minute into the crisis. And so they went from tens of thousands of feet flying through the sky and ended up spiraling into the ocean, tragically losing all aboard."
http://www.econtalk.org/archives/2015/11/david_mindell_o.htm...
A configuration management system that incorporated spaced repetition would be cool. Every once in a while, go into an incremental mode where you actually type in the commands. It has the added bonus of getting new people aware of the system. Sure, you can always just go read the source and figure it out. Having the system force some human awareness from time to time would be handy.
Have them both handling control at the same time, reconcile the inputs in a sane manner (or have a master/slave where one is providing phantom inputs). Added benefit of being able to error check each other.
(Supposedly the plane is supposed to loudly complain if the inputs diverge. I'm not sure if this happened.)
In flight testing, whenever we encounter situations like this (the aircraft does not react as the pilot expects for a given input), we don't necessarily blame the pilot. Especially if multiple pilots get into the same mess. Then we know it's time to change how the aircraft behaves and bring it in line with what the pilots expect.
In fact by the time it arose, the aircraft may already have been in an unrecoverable dive (or whatever the word is for rapid descent while stalled), I don't remember. So it's not a primary cause of the accident.
There's no substitute for physical feedback. It's instantaneous and impossible to miss. It's completely beyond my comprehension that Airbus eliminated it.
As I recall, the one pilot held full back stick all the way down, while the other pilot tried and failed to recover. Releasing the back stick and executing a normal stall recovery would have saved the airplane at just about any time.
As for your last point, I'm not an expert, and I'm not as incredulous about this scenario as you seem to be. (As a teenager I did fly a lot in PC flight simulators, though.)
The non-technical Vanity Fair article I referred to claims that the accident investigators estimated the last point at which the aircraft could have recovered from the dive was around the time it passed 13,000 feet:
"Though precise modeling was never pursued, the investigators later estimated that this was the last moment, as the airplane dropped through 13,000 feet, when a recovery would theoretically have been possible. The maneuver would have required a perfect pilot to lower the nose at least 30 degrees below the horizon and dive into the descent, accepting a huge altitude loss in order to accelerate to a flying angle of attack, and then rounding out of the dive just above the waves, pulling up with sufficient vigor to keep from exceeding the airplane’s speed limit, yet not so violently as to cause a structural failure. There are perhaps a handful of pilots in the world who might have succeeded, but this Air France crew was not among them. There is an old truth in aviation that the reasons you get into trouble become the reasons you don’t get out of it."
When trying to get the plane onto the center of the runway, both pilots' actions are probably pretty correlated, and averaging could remove any noise (e.g., twitches).
But in many other circumstances, it seems like it would be downright dangerous. Something's in front of the plane. One pilot goes to the left, the other pulls to the right, and the system averages the commands and smashes straight into it.
The benefits of a marginally straighter landing don't seem like they would outweigh the potential of a massive catastrophe.
Reviewing changes by hand every few months keeps you aware of what needs to happen. Automation doesn't get brittle, people forget how systems are flexible and inflexible. Spaced repetition would give people a chance to keep up with the current design, so when that crazy security vulnerability happens, you can jump in and change stuff knowing how it works, rather than having to figure it out on the fly.
I don't think it would take much. Configure a box today, tomorrow, next week, next month, 3 months, then perhaps every 6 months.
There are very smart people that sort of intuitively keep up with those kinds of changes. But those people change jobs. How do you get a mere mortal up to speed on 20-30 machine configurations? config management can configure hundreds of machines a day, no problem. but it's easy to forget what's really going into each of those boxes.
This is more about preserving organizational awareness, not so much robustness of a running system.
http://www.vanityfair.com/news/business/2014/10/air-france-f...
THE MACHINE STOPS
US military systems pervasively use inertial navigation (INS), many of which can accept micro-corrections from the GPS system. However, they only accept corrections within the intrinsic error bound of INS, which is quite small. Spoofing GPS can buy you a deflection of meters at most, not kilometers.
The US military is developing new types of inertial navigation systems that are so accurate that it obviates the use of GPS altogether. (And the media will probably still call those systems "GPS-guided".)
The INS system you mention has small errors, which aggregate/accumulate over time, this time being in minutes and tens of minutes.
These errors cannot be rectified inside the INS, from what I understand. They require external fixes, like GPS.
I am willing to be wrong however.
There are a couple things that people forget about US military INS:
US military INS actually operates as a giant swarm of INS computers with other sensor inputs that compare notes -- wisdom of crowds -- to correct accuracy. If one INS computer drifts too much, the other INS computers will notice and correct it. Even without the swarm, many larger military systems will have multiple INS computers distributed throughout the platform that can compare notes.
The US military has been at the forefront of inventing exotic INS technology since the 1960s, and the details of their INS capabilities are a closely guarded secret. It is used ubiquitously in almost every combat system they produce. What is known is that even in the 1990s the accuracy of INS significantly exceeded design requirements, and there are current military research programs that suggest they know how to build INS that exceeds current GPS accuracy over long temporal baselines.
Given that they already have a GPS constellation that they use for a wide variety of other cases, GPS-corrected INS guidance is a very cheap way to squeeze some additional precision out of weapons that were already pretty precise in the first place. It allows them to build a dirt cheap and reliable guidance package that is also highly precise in environments where GPS spoofing is not a credible problem. Remember, these particular guidance packages tend to be modular and swappable: use the cheap stuff that leans on GPS more for low tech enemies, use more sophisticated and expensive INS for the high tech enemies.
Certain systems do indeed operate as you say, with INS as the primary guidance/navigation mechanism. Submarines for instance, cannot rely on a GPS fix. Strategic missiles as well.
But ground forces do indeed use GPS systems. INS systems are very costly to build and maintain, go GPS is a go-to for missions and requirements where it's a good fit. You simply cannot outfit every ground-pounder with an INS system and then keep it properly operational in their task envelope.
Also, parts of the GPS system are quite jam-resistant. Not entirely, but when it was first developed, GPS was pretty secure for it's time, using a cryptographically derived spreading code for the carrier.
Here is an example of a system in use today:
https://en.m.wikipedia.org/wiki/Defense_Advanced_GPS_Receive...
Some of you might be interested in an article I wrote on the subject:
https://sofrep.com/46818/gps-spoofing-how-iran-tricked-us-pa...
[1] https://en.wikipedia.org/wiki/Joint_Direct_Attack_Munition [2] https://en.wikipedia.org/wiki/Tomahawk_(missile)
Along the same lines:
I work in manufacturing, and design systems to automate processes. "What is the current, manual process?" is the first thing I ask for when beginning a new project. I then try to design the automated system, so that the manual process can be used as a fallback should any part of the automated system fail. What happens when we have a network outage or something similar? The people involved suddenly forget the manual processes that they had used for years, and sit around twiddling their thumbs and calling IT every 5 minutes asking when the system will be back up.
What Iran did probably simple jamming that caused drone to either crash or emergency land.
"... Iranian engineer's assertion that the drone was captured by jamming both satellite and land-originated control signals to the UAV, followed up by a GPS spoofing attack that fed the UAV false GPS data to make it land in Iran ..."
Wikipedia has more of the story https://en.wikipedia.org/wiki/Iran%E2%80%93U.S._RQ-170_incid...
That was disputed "... GPS is not the primary navigation sensor for the RQ-170..."
Anyway, thanks!
In recent years GPS modernization means satellites have been using new codes on L2 (L2C) which are un-encrypted. Only about 18 of the satellites carry that capability at this time.
Cruising at altitude is not that hard. I'd be far more concerned if they were interfering with the various Instrument Landing Systems during inclement weather.
> Every, say hour you do it manually buys you a few hours of autopilot.
Most pilots already hand-fly in clear weather during landing and takeoff by default; somewhat for the reasons you suggest, but it's also the safer as it actively maintains pilot situational awareness.
You're generally only going to use the A/P during cruise, departure and approach or during instrument flight conditions.
Not in RVSM airspace above FL290 (29,000 ft). In RVSM [0] airspace its required to have a functioning autopilot coupled to an altimeter of a certified accuracy (+- 65 feet)
Most pilots couldn't fly the required accuracy, that high up, for any length of time, with manual hand-flying. Even in a FBW airliner, it would be hard.
[0] https://en.wikipedia.org/wiki/Reduced_vertical_separation_mi...
https://en.wikipedia.org/wiki/Coffin_corner_(aerodynamics)
Being near the coffin corner is often desirable for performance reasons, but it sure isn't easy.
Of course, GPS is a rather old system and I'm unclear how you would add encryption in a way that it's not compromised when someone gets access to a decoder but also doesn't endanger "mission success" because a encryption key punch card is missing and the drone refuses to start. It's like the nuclear launch codes being all zeroes.
The keys are rotated periodically, so if somebody stole a decoder, it would only work until the key in memory expires. (Military GPS receivers also have a "zeroize" button to destroy the keys in case of imminent capture.)
Another fancy way to navigate without GPS is to use automated celestial navigation. The SR-71 had one of these in the 1960s, and it's also good for submarine-launched nuclear missiles. The hardware is able to sight stars even in the middle of the day (and not just the Sun, funny guy).
For commercial aviation, the typical backups are radio beacons such as VOR and NDB, inertial navigation systems, and good old dead reckoning plus pilotage (i.e. looking out the window).
Also with the reliance on GPS and its inherent fragility has caused the US Navy to restart training in celestial navigation. In order to reboot the training regimen, the Navy is relying on Coast Guard instructors, since the USCG never stopped.[1]
[0] http://www.airliners.net/aviation-forums/tech_ops/read.main/... [1] http://www.militarytimes.com/story/military/tech/2015/11/01/...
I recently spoke to old aircrew guys I worked with and apparently they stopped teaching manual celestial nav in the late 90s finally. Automated celestrial nav was, per my understanding, never too common although SR-71 aircrews commonly used its system.
A variation on terrestrial celestial navigation was used to help orient the Apollo spacecraft en route to and from the Moon. To this day, space missions, such as the Mars Exploration Rover use star trackers to determine the attitude of the spacecraft.
> Another fancy way to navigate without GPS is to use automated celestial navigation.
I figured as much. I'm just under-informed in this area, and try not to state things as fact that I don't know as such. I did mention the stars, but it didn't occur to me they are visible during the day with the right equipment. :)
> inertial navigation systems
I'm aware these exist (due to some military fiction I've read), but that's the extent of my knowledge. I'm not aware of how accurate they are.
> good old dead reckoning plus pilotage (i.e. looking out the window).
I was thinking of systems that replace pilots, even if for short whiles, not supplement them, so discounted human correction while in flight.
As far as quantifying that growing error, Wikipedia says it's typically less than 0.6 nautical miles per hour. An airliner after a long oceanic flight could know where it was to within a few miles, good enough to reorient and find the destination airport.
Early efforts in autonomous navigation came out of a strong desire to blow up the Soviet Union, so cruise missiles and ICBMs and such are a good place to look if you're interested in early examples.
Source: Worked on said GPS/INS systems in the military.
They "require" GPS input in the sense that if I handed you an INS that was initialized/calibrated with some precision (Oministar/RTK/P-code GPS/ whatever) reference but had been running in the trunk of my car (or dangling from the collar of a feral cat) for N hours/days, there's no way you'd fly a plane with it without initializing it with some sort of position referene.
(If you came up with "sit it on some previously surveyed datum, and you don't need GPS" on your own, you get an A)
It used to be that the clear channel was less accurate (even beyond the 'selective availability'), but signal theory got us to cm granularity without the computational expense of the military channel.
Just sign the signals with RSA. This depends on keeping the private key out of hostile hands, or allow key revocation and generation. But it's possible in theory.
Also, i'm not so sure about the graceful failure. Hypothetically, the human takes over. But if the human hasn't actually flown in months or years the'll likely be kind of rusty. Now you're throwing them into a complex situation - the autopilot can handle the simple stuff. Coupling weak skills with difficult situations seems like a bad idea.
I kinda think autopilots and self driving cars should give a limited clock. Every, say hour you do it manually buys you a few hours of autopilot. Just to keep skills relatively fresh.