Redeyz

An engineering grad and I were discussing the likelihood of a multi-engine transport drone/airplane.

We are both airline pilots that fly the 757/767. Obviously, this airplane would never be retrofitted in large numbers of aircraft to be an operating drone.

What we focused on was not the theoretical likely hood that a multi-engine drone could be produced. But the "practical" likely hood that they could reliably remove the human element from such a complex task of operating an airplane in the National Airpace Structure.

During the flight that this discussion took place, we were departing Newark, NJ. As you might know, we are limited to 250kts, by regulation, under 10,000 ft. The particular mode that we needed to use for the departure out of Newark, the airspeed was limited (by the mode we selected) was 215kts. However, after takeoff, the flight guidance system that should have limited the airspeed to 215 kits faulted and was accelerating towards 250kts. We intervened and corrected the speed problem manually, which is not uncommon to have to correct from the autopilot/flight guidance systems.

This and similar incidents happen on nearly every flight (that require human intervention) was the genesis of this topic. How could a drone reliably be created to operate in our system when it can't handle small speed or altitude constraints. How would a multi-engine drone ever by able to RELIABLY handle an engine failure (V-1 cut) when 2 appropriately rated and trained pilots are working quite hard in the simulator to handle such complex emergencies.

Practically, I feel we are a long way from that day where a pilotless multiengine drone can safely and reliably fly in our system, without human intervention.

But, I'd like to hear other views.

Cubdriver

Money is the only question, not whether it is technologically possible. The only question after that is will the funds be raised due to war-time needs, or will they be raised by commercial industry. We will probably see remotely piloted cargo UAVs before we see fully autonomous cargo drones because they are cheaper. Fully autonomous drones will start with small short range tasks like the Amazon item and when the issues are sorted out, larger autonomous drones will gradually come online. It will be cost driven as industry sees the money in doing it. The Amazon drone is probably several years out due to regulation and safety concerns, but the guidance technology for it already exists.

As for the particular failure scenarios you mention, we already have flight governance systems that can do anything imaginable with an airplane or helicopter at least while the guidance system is working properly. The million dollar question is, how much more R&D money will it take to make these systems as reliable as trained flight crews. No one knows the answer to that now, but as mentioned large leaps in technology are war-driven and barring that, it will take industry a long time to see the savings and a good enough reason to invest the money.

JamesNoBrakes

The russians launched a space shuttle into space, it orbited the earth, re-entered the atmosphere enduring a crazy amount of stress and heat, and landed like an airplane, all with automation and remote control...and that was 25 years ago.

Heck, just think about your autobrakes...lol, they took braking away from humans a long time ago!

While there are numerous problems to be overcome with automation and integration into previous architecture and systems, the indescribable march is impossible to ignore IMO.

There are many ways that one could have "fail-safes" designed into such a system, sensors that determine if the systems are operating correctly, and if not, they choose a different mode of control. AI is becoming extremely advanced and capable. At some point, even though there will still be equipment failures and malfunctions, the amount of failures from a completely automated system will become lower than one that requires human interaction. That day is coming, whether you like it or not and it's already come in many industries.

Luckily we don't really have to worry about this impacting our jobs as pilots in our lifetimes, although I'd say we are seeing effects in society as more and more people are displaced from traditionally required jobs.

Busdrivr

How do you fail safe spoofing?

Redeyz

Part of this question, that I did not communicate very well, is that the problems that we face on a daily basis are because of the intense congestion with our airspace system. So, the one off Automated Russian Rocket does not cut the mustard in my estimation. Primarily, because there won't be 1 drone flying per day in restricted airspace away from populated areas. As we, and 1000's, of airplanes take off and land in Newark, JFK, La Guardia, and Teterboro, it only takes 1 mistake to have 2 drones raining down on densely populated areas below.

So, when I originally asked the question and said, not the theoretical, but the PRACTICAL, this was what I was referring to. My apologies for not painting the operating atmosphere properly, hence the tolerance for error much less.

CRM114

Saving the day with speed intervene is pretty common when your 3 generation old FMC technology craps out. The hardware/software combo was STC'd in the 80's which makes it about as good as what the Space Shuttle had (512K of RAM).

You guys are engineers, take a look at the hardware/software you operate and apply Moore's law, my iPhone looks like a supercomputer compared to what you're flying. It will happen, the only unknown is when.

Cubdriver

It's like an automated assembly line in 3 dimensions operated by robots. Humans monitor from distant points and decide a few issues but most of it is automated. These are already very safe systems and they have been in use for decades in manufacturing. The more difficult challenge will be integrating our automated (drone) aircraft with old-school human-driven craft, and the latter must stay in place for the forseeable future. It is a difficult challenge because people do not like being driven around by machines and probably never will, even though in time the drones will be equally safe. And if human pilots must be used for some aircraft, it becomes difficult to integrate the two types of airborne vehicle.

It is easy to predict a two-tiered system will evolve where drones do everything on one side of an airport while the second side operates where (for a price) you can still get human pilots. The two systems will be separate in which runways they use, but the difference may be as small as having separate runways for the two users. We are already rapidly getting the point of automated ground control at airports now and it is not a stretch to see drone traffic working alongside human aircraft. I tend to think however that some degree of separation will remain for a very long time, we will probably not see total integration during our lifetimes nor those of our children.

Grumble

A third world, African country just booted the US Air Force because they were tired of them crashing small, rotax powered UAV's in their back yards. This from a program that has had 20 years to mature from billions of dollars, and a not-for-profit organization.

JamesNoBrakes

Yeah, ok, while the technology is being used all over the globe and we have self driving cars...

machophil

could a computer safely deal with all forms of hazardous weather?

could a computer still fly after a major system malfunction?

could a computer safely deal with all forms of hazardous weather after a major system malfunction?

could a computer make a sound aeronautical decision based on information that does not conform to it's programming?

could a computer pull off the "miracle on the hudson"?

could a computer be hacked by an evil-doer?