Germanwings plane crashes in French alps with 150 on board

Agreed.

 

A more logical solution is to have the computer override human operators from making a fatal error, with the only way to disable the computer control is to require a biometric scan from a high ranking airline official on the ground to input an override code.

 

They have a system in place that prevents a lot of crashs (no rapid descending possible etc), but it's hard to do this. Slowly descending is not alarming to a computer without proper context, the only thing preventing a crash would be if the system analyzed the whole terrain in sight. But even then, preventing a crash into the mountains would be nearly impossible given the speed of a plane.

 

Sensor technology has advanced rapidly within the past few years and microprocessors within the cockpit should be able to give readings on all variables taken into consideration and determine that if descent is not necessary it should not be allowed.

Computers will be driving our cars for us within 10 years, they should be flying our planes now.

 

If the computer is making a fatal error due to a weird situation, time is critical - the time spent trying to expalin to your ground official what's going on and disable computer control could lead to its own disaster.

The reason that computers aren't driving all our cars now is that there are still way too many variables and things that could go wrong. The technology is simply not there yet outside of limited test environments.

 

True. That's why the basic rule is "Aviate, Navigate, Communicate" (in that order)

Correct.

 

That is not quite correct. The technology is there, but the logistics aren't. Implementation and integration are the major hurdles.

Computer error is far less likely than human error.

Google estimates 10 years for mass integration but the technology is mostly already in place, the final stages of testing are currently being worked on which involve variable control like weather conditions and if or when to allow human intervention.

Self-driving cars have already made coast to coast trips across the states. Once the Internet of Things becomes more mainstream and computers are able to take readings from embedded microprocessors (like a sensor on a bridge in icy conditions) then full integration will be easier. But admittedly it will still take quite a while for full-scale implementation due to a number of factors (bureaucracy, corporate interest, etc).

 

You should probably read this:

http://avherald.com/h?article=47d74074/0000

 

Looks like a clear-cut case of stupid engineering. As the first comment pointed out, a frozen sensor has no variance and should be easily detected by a simple software patch.

 

You could put a third guy in the cockpit. That way, even when one guy is in the john, there's never just one pilot in the cockpit. For a pilot to crash a plane, he'd have to either enter a conspiracy or win a fight. Which is doable, but should cut down the frequency of something that is already very rare.

 

This story keeps on getting more tragic.

As far as being able to fully automate the flight as others have said I don't think the technology is there yet. I think you will still need a rapid override to be able to deal with unexpected situations like a bird strike. The other thing I wonder about with automating the whole flight is given all of the issues with security is how much security will be needed to prevent flight systems from being hacked. Terrorists in the future might not even need to bother to hijack and fly a plane itself into a building but could hack it.

 

Good idea in theory, but not very cost effective to have a third pilot on board. Like you said these events are very rare to begin with so I don't think any change will come about.

 

Info on what is known about the flight's descent from the BBC:

http://www.bbc.com/news/world-europe-32072218

It seems it was definitely intentional. Guy waited for pilot to leave, made his last routine contact, and then punched "100 ft" into the requested altitude. Never uttered another word in his miserable, horrific life.

Prayers to the victims and their families. At least the end was instantaneous death, since they hit the mountains at full speed.

 

No matter how advanced technology gets, I guarantee you there will always be a scenario computers won't understand how to handle, or there will be some kind of computer failure it can't recover from. The second you have a plane full of people who die because the perfectly lucid, skilled pilot wasn't allowed to override something stupid the computer was doing (or not doing), we would be right back where we are today, with overrides and the possibility of some idiot/terrorist/suicidal maniac capable of killing people. As others have said, there will never be a perfect system.

 

Then you could simply require both the pilot and copilot to enter their own override codes at the same time to take over manual control if necessary.

You guys may think the technology isn't there yet but it is. Google has driven their driverless car 400,000 miles without an accident.

Airplane autopilot systems have been in place since the 80s.

The problem is the technology still requires sophisticated and fail-proof programming designed by humans.

 

I agree with Major. A remote override is a big mistake. First, every tower in the flight path would have to have this "key." Second, all "keys" can be broken, *especially* remote computer keys. Third, how would you truly know the real pilot is making this request. Do you give him a password, that a terrorist could torture out of him?

This change would actually increase the risk of that door being opened by terrorists. Computer Software security and hacking is part of my job. I would never recommend a remote "backdoor" override.

 

Crashes like this cause me to believe that at some point, perhaps hundreds of years from now, there will be absolutely no human input on airline travel.

Even now, you can fly an entire flight and only control the plane during takeoff/taxi. Some airports are even equipped with "autoland" systems so that there is absolutely no human input.

I envision a time where there aren't even cockpits in the planes, they just fly autonomously, onboard there is only a cabin with flight attendants, no pilots whatsoever.

It'd be a much safer world.

 

People are talking about autonomous planes, manual overrides, remote overrides, and hacking -- big capex dollars and new risks. A third guy in the cockpit eliminates some of the same risks these other solutions address for a little additional operating expense. If an airline has to pay $150k/year for that pilot, and he flies 300 flights a year (both conservative estimates, imo), he adds an incremental cost of $500/flight. If each flight has 180 seats filled (a 737 has 189 seats, I believe), it'd add about $3 per ticket to have a third pilot. That's $3 wasted, probably, since we're mitigating something that happens once in a decade maybe. But the union would be pleased.

 

Such an idea is not only good, but practical as well. In fact it is already implemented in the U.S. This tragedy could've been prevented has Europe implemented the same measure.

http://news.yahoo.com/official-1-pilot-locked-crash-plane-cockpit-072049706.html

 

I didn't know that.