What would have happened if the automation had said to itself, "hummm, the airspeed sensors seem to be taking a break, let's ignore them for a while and see what happens if we just keep doing what we have been doing for the last 3 minutes...."? May be flash an ECAM warning stating they were in "speed extrapolation mode" or something?
This is extremely valid point, one that all technical guys should be discussing.
There are a variety of new, very cheap sensors that could be added to every plane to make something like that possible. In this case, equipping the plane with multi-axis acceleration sensors would allow very good approximation of air-speed for hours even with all pitot tubes out of commission. These solid-state sensors have been around for almost 20 years and today everyone has one of them inside their smart phone. They are very inexpensive and extremely accurate.
There is a certain amount of idiocy associated with the way automation is pursued in all "regulated" technology areas. Nobody dares to say that the only way to safely do something like that is either:
1. Full sensors and systems redundancy where each reading needed for proper functioning of controls is replicated 4 or more times, leading to the situation where complete loss of necessary control input would require several unrelated but simultaneous failures.
2. In absence of full sensor redundancy where dropping of automation due to unreliable or missing inputs is allowed as even fairly unlikely event, the crews should be able and required to regularly practice flying without the automation.
Sensor redundancy is possible but is not done due to regulatory costs. To begin with, Pitot Tube is an archaic piece of sensor that should not be used anymore. It is being used because introducing new technology requires decades of bureaucratic and legal hassle. Not only it has to be approved by variety of bureaucrats but anything new introduced to the market (no matter how good) is instantly a target for lawyers. If the AF447 had brand new kind of any type of sensors the lawyers would instantly claim that this was the reason for the crash.
To see how logic and sense have nothing to do with today's law, examine all the lawsuits against Toyota, for example - where people claimed that their car accelerated on their own. The only way to end this nonsense would be to allow libel suits and damages against parties that make such idiotic claims.
In airplane market - look how many years it took them to start using solid state voice and data recorders in place of the mechanical ones with magnetic tape. Here, the decision to use them should be crystal-clear.
1. Even most retarded lawyer cannot claim that using solid state recorder instead of magnetic tape can cause a crash
2. Solid state recorders are much smaller, able to withstand moisture and heat that no magnetic tape can.
3. Solid state recorders are much cheaper. For the price of one mechanical recorder you can have ten solid state recorders, one in every part of the plane, each recording a full copy of the data.
The problem is that while there is a big demand for automation (mostly because of fuel efficiency concerns), the regulatory framework is absolutely not able to allow to do it right (the full redundancy path), and so it is done same way if someone was trying to put a fuel efficiency automation on a stage coach.
With such approach - it makes perfect sense to drop automation due to unreliable sensor reading because in absence of sensor redundancy the pilot still has more sensors than the plane and in theory should be able to handle the situation - IF he is trained to do so, that is and can use his eyes and ears as a substitute for a faulty sensor.
Pitot tube - Wikipedia, the free encyclopedia
Modern ways to very accurately measure air-flow:
Mass flow sensor - Wikipedia, the free encyclopedia
Airplane application of such a sensor would basically require a little metal peg or even simply piece of metal foil stuck on a wing. The principle of operation is such that you heat a piece of metal to a specific temperature and measure the amount of energy required to keep it at that temperature. The bigger air flow cools it faster so more energy is required to keep it at constant temperature. Damaging/disrupting/contaminating such a sensor is orders of magnitude less likely than Pitot tube.