jtbost

Exactly....

tallpilot

With all due respect, every head in the sand 'automation ain't taking my job' argument assumes that if the computer can't figure it out the airplane just runs out of fuel and crashes. These flights will be constantly monitored from the SoC and there will be multiple ways to upload changes including taking over full control manually.

Do you think it might occur to them to put a video camera on the nose to 'see' that building cumulus you mentioned? Optical sensors keep improving and so do the algorithms used to analyze the data. I don't mean to be insulting but it is truly inconceivable to you that engineers can't figure out how to get a computer to steer around a cloud??

First we will fly these things single pilot to monitor the automation to see how it does. Then we'll move to being on standby in the SoC to take over manual control, then they will figure out they don't need 'real' pilots for that.

I know nobody wants to think of the day the skills we've spent a lifetime perfecting are nearly worthless but ignoring trends and pretending it won't happen is irrational. The only question is can we make it to retirement before it does.

USMCFLYR

Countless false localizer signals?

Not in the US.

ReadOnly7

It only takes one, and I’ve certainly encountered more than one in the “perfect” system we have here in the US.

AllYourBaseAreB

A seeming design flaw in my last type lead to a lot of false captures when captain’s side was flying and had ILS capture armed when in magenta. The whole fleet.

pangolin

You haven’t flown much in the northeast then.

pangolin

There is more. A lot more. I have issues with the whole text stuff to the airplane path we are on. I build situation awareness listening to the radio. Maybe comes from teaching in SoCal airspace pre tcas. At any rate I can “foresee” issues and problems by listening. I’ve told FO’s This or that’s going to happen. Stay up here a bit. We are following a heavy that they didn’t tell us about etc etc. This ability is lost on automation.

Setopbug

My airplane's crosswind limit is 25 knots. My limit is 45 knots.

It's going to be decades, and will require a clean sheet aircraft and systems architecture design, and a complete change in how the NAS is run. And you've probably seen how quickly the FAA embraces, procures and refines new technology.

Optical sensors are improving? LOL. The USAF can't deliver a FMC refueler based on existing airframe technology after two decades.

Relax. Your concern should be how you're going to keep from having to spend your life savings on medical care. Nearest, hottest threat, buddy.

bradthepilot

How many years do you have building complex software systems? Integrating that with hardware? Doing a security analysis including pen testing? Bluntly, I have way more experience in all those fields than you do, and I can promise you that making the computer "figure it out" is the easiest part of all that.

Proving that it isn't susceptible to black-hat actors is pretty darn hard. Proving it works in a virtually unlimited set of scenarios is harder still. And if there is any "artificial intelligence" involved, proving its functionality approaches the impossible.

In technology, it's often the validation ("testing") rather than the development that tends to drive costs for large scale projects, so this will be an extremely expensive undertaking. Has there been a cost vs. benefit analysis done to show that such an undertaking is worthwhile?

These all come together to form a pretty complex system. Taking just a thin slice of functionality, you say that the aircraft could be controlled directly from the SOC. How are they authenticated (i.e. how does the A/C know it isn't connected to an imposter SOC)? How redundant is the SOC? Where are they (you weren't thinking it was just a single location, were you?) located? How do they talk to each other? What is the failover protocol? How is a compromised SOC detected and locked out? What is the data communications protocol and how much bandwidth will be required?

If you think this all happens quickly, I refer you to ADS-B, which is a much simpler system that took decades to materialize. So no, the "head in the sand" folks as you call them, are correct. I have no doubt that such a system will eventually happen, but there is nothing to indicate that it will happen before any current pt. 121 pilot retires.

Downtime

737 will start descending on a glide slope before capturing the loc if you don’t wait to arm app mode. I have lost count of how many they have reminded us in training.