Airbus Defence Sagitta (small UCAV demonstrator)

Reaper

ACCESS: Secret
Joined
31 March 2009
Messages
271
Reaction score
78
Airbus Defence flew for the first time the Sagitta UAV in South Africa. It's a collaboration project from Airbus and different universities.

http://www.airbus.com/newsroom/press-releases/en/2017/07/Sagitta.html
 
The original rendering for this project shows it without vertical fins. Do we know if they will remove them later in the demonstration or have they gone with a "safer" design?

I had thought Sagitta was mainly about tailless flight control demonstration (alongside the other cooperative effort).
 
Hi,

The flying public may not be so interested in it—yet—but pilotless commercial airplanes could be
a boon to the aerospace industry and some financial analysts see a natural evolution in that direction coming. “Meaningful savings can be generated via mission optimization, greater predictability and
reduced flight crew and training costs,” a new report from UBS aerospace, airlines and logistics
sector analysts. “Younger (18-34) and more educated respondents were.

http://aviationweek.com/commercial-aviation/pilotless-commercial-aircraft-follow-money?NL=AW-05&Issue=AW-05_20170808_AW-05_938&sfvc4enews=42&cl=article_1&utm_rid=CPEN1000002229670&utm_campaign=11212&utm_medium=email&elq2=193216aed2e34db3820e1811fdb40540
 

Attachments

  • sagitta.jpg
    sagitta.jpg
    39.1 KB · Views: 250
@hesham

offtopic: I feel the drone trend might have reached its peak.
 
Reaper said:
@hesham

offtopic: I feel the drone trend might have reached its peak.

Hi Reaper,

but it's very hard to depend on a computer to drive commercial airplane,the passengers
will not feel safety.
 
Ok I just read the article, which talks about passenger acceptance of unmanned aircraft, which is unrelated to Sagitta.

I think when people can get used to self driving cars, why not to self flying planes.
 
Peoples can get use to things noboby thought it could be possible.
Me think the problem with drones carrying passenger would be safety regulations.
An AI or remote controlled machine can work in common situations , usual take off, landing , cruise... Ect
Airliners already cruise on autopilot most of the time, and can land on auto too nowadays. So it's almost like drones with pilots monitoring on board.
Problem is that some flight incidents can be results of very complex and weird situations that for now I don't think an AI could resolve. Sometime the situation is so messy that only a human could save the passengers, or not... And if remote controlled, you have to make sure the connection is very very safe.
So before regulation agrees that an remote or AI controlled airliner can be used without pilots onboard, it will take some time, IMHO.
 
yes. This is exactly what I was advocating in the Stuntman thread on Keypub/civil aviation*. Fact is that many situations are not predictable especially with the change at pace in the world (climate change that impact turbulences, weather event, the impact of ever increasing electronics, flooded airspace with with harder interferences (generalisation of powerful electronics etc...) The unknown are too numerous while the tempo and complexity of changes increases.
Still, it is perfectly concevable that a large portion of flight activity could now be handled both, remotely and be automated, leaving the pilot dealing only with all the potentially hazardous situations. My motto was to have a single onboard pilot dedicated to flight safety, a backup flight intendant with increased responsability over passengers (take-over all the ones usually handled today by the flight Captain + all the extras that the mass-air transportation incur) and an offboard team of specialists in charge of the normal flight event, from flight planning, boarding, cruise conditions, flying or communication etc...

On economical grounds, the advantages are numerous. The offboard team would be able to handle year after year even more flight thanks to the help of artificial intelligence and the onboard pilot would gain a tremendous experience in a shorter time. For that later point, it is easy to understand that to let the pilot pro-efficient and his postion kept valid on economical grounds, he would have to be kept active through in-flight training encompassing real-time piloting (or virtual stick handling where his performances would be evaluated in real-time Vs the offboard pilot or an IA for example ) and would fly way more often.

All this is explained more in the details inside the Stuntman* thread on keypub forum. Feel free to participate with your inputs.

A pilot should now be an incident specialist. A manager of catastrophic failures. A trapezist with skills selected and refined everyday trough such a hard training. Someone able to survive (with all of us behind) the most daring flight situations when it may occurs.

Is that a commune acceptance now among the airliner industry that the formations are too weak, too uncertain to let the pilots fly the plane even when all the systems are down ? No.
If a plane can't be manned, it simply never have to be manned. [...] Pilots are not here to be only the warranty of an intermediate software release.
[...]

The stuntman could have save[d] them. The hundreds of them...

*from "Let's bring back the stuntman": http://forum.keypublishing.com/showthread.php?133640-Let-s-bring-back-the-Stuntman
 
I think the main problem with human-overseen AI-controlled vehicles (cars as well as aircraft) is the longer reaction time humans have. The driver/pilot is likely to be doing other things (i.e. not constantly looking out and being aware of their surrounding while the system is operating correctly but when a problem occurs and the AI shunts back to manual mode you are not paying full attention and suddenly have to assimilate what the problem is, what is causing it and assess the risks and maneuvere accordingly to avoid the situation the AI couldn't handle.
Admittedly the road traffic situation is probably more confusing and chaotic and the reaction times have to be quicker to take back control than most aircraft flight situations but as Air France Flight 447 showed, humans can interpret data incorrectly when overwhelmed with data without visual references. Would an AI pilot have acted differently in that situation had it been able to react to unexpected scenarios?
 
Very interesting topic, though we are OT in regard the Sagitta I think :)
Are airlines, insurance Co, flight authorities… ect ready to leave the sole responsibility of say 250/300 passengers to an AI or a remote controller…
Even though a human pilot is fallible, there is still the problem of responsibility.
There still need to be someone "in charge" on board, even if only for monitoring.
It’s a problem for Mr.Asimov :)
Though if it’s a way to make more money, it will eventually happen, but not very soon.
Also you need to have a crew. 250/300 peoples (or even less) it’s a crowd, and many still needs to be guided to their seats, feed , solve human behavior problems and all that.
 
Something ate yesterday's reply (I probably pressed back instead of send), but personally I'd prefer to have a pilot aboard.

And I say that having been on both the 777 and Typhoon FCS development teams. The FCS is there for the long hours of mundane long-haul flying, but for the stark, screaming seconds of imminent disaster, I'd prefer a human in the loop.

You might be able to spin passenger acceptance out of driverless cars, but the problem space is very different. A car control system always has the option to stop in the event of a problem, that doesn't work quite so well at FL30 over the Mid-Atlantic. There are issues with having a pilot aboard, I know the accident stats, but those call for better training and work on the pilot-cockpit interface (including the pilot's conception of that interface), not doing away with the pilot. An automated FCS can deal very well with issues which fall within the scope of its programming, which prospectively can include adaptive reaction to failures, but it remains a single purpose device bound by its programming, while a pilot is a general problem-solving intelligence, able to make it up as they go, and that's the kind of resilience I want at the pointy-end of the aircraft.
 
Hood said:
but as Air France Flight 447 showed, humans can interpret data incorrectly when overwhelmed with data without visual references. Would an AI pilot have acted differently in that situation had it been able to react to unexpected scenarios?

In AF447, the on-board AI - the flight control system- shunted control back to the pilot specifically because it had no air data to fly the aircraft with because of icing of the static ports. I do actually think there is some potential for the FCS to handle such a situation, but there's one system in the cockpit that's fitted with a completely independent inertial platform and the ability to estimate missing air data - the pilot.

The problem in AF447 was the pilot flying fell victim to a false perception of what was happening, a somatogravic illusion, and kept the aircraft in a deep stall all the way to the ground, shouting down the more senior members of the flight crew when they tried to get ahead of the problem. That points to two training issues, upset recovery and cockpit resource management, and possibly a psychological issue around aircrew screening, plus the need to better signal FCS failure states and mode changes to the flight crew, but it doesn't point to getting rid of the pilot as a solution because the accident was triggered by a complete air data failure, and the pilot is the only solution we have to that scenario.
 
DWG said:
but it doesn't point to getting rid of the pilot as a solution because the accident was triggered by a complete air data failure, and the pilot is the only solution we have to that scenario.

Just fly off inertial inputs rather than air data and restrict the flight envelope accordingly if needed.

Air data fails -> gives detectable failure modes e.g. out of range, rapid changes etc. -> declare air data inputs as bad and revert to inertial -> restrict flight envelope and control inputs if necessary -> recycle air data system to attempt to repair
 
red admiral said:
Just fly off inertial inputs rather than air data and restrict the flight envelope accordingly if needed.

That's what I'd advocate to get the aircraft through a short term loss of data, we use rolling averages for everything anyway, but I have my doubts about it on a long range flight, or in the face of a major upset. Equally a pilot can function in the face of the loss of both air data and INS, an FCS can't.
 
TomcatViP said:
an offboard team of specialists in charge of the normal flight event, from flight planning, boarding, cruise conditions, flying or communication etc...

Separating flight planning and piloting is just plain dangerous. It means that when control reverts to the pilot they aren't aware of active NOTAMs and other conditions the planners have checked on, such as meteorological expectations. Removing comms will worsen that, because he won't be aware of surrounding traffic. The pilot's job, and the stuff he must retain a background awareness of, extends beyond the aircraft. Equally the planners can get it wrong in ways the pilot needs to be aware of. There have been several serious failures in fuel planning/delivery, the Gimli Glider was one IIRC, but if the pilot hasn't been in charge of the planning, you lose the chance for him to say "we're handling lighter than we should be, contact the tower and get them to check how much fuel they put aboard."

The fact is the old barnstormers had an appalling accident rate, and the professionalism in the way we fly now emerged from understanding what they were doing wrong, and systematizing the processes to squeeze error out. And part of that systemization is having the pilot do the planning because they are in the best place to catch errors emerging from it, and they need to know all of that information anyway.
 
You are absolutely right. But that was not my meaning ;)
In no way would I say that everything is done wrong etc.. etc... The focus on procedure and the need to alleviate the necessity of being a "system specialist" to fly some of the planes around at the detriment of basic flying qualities was the inspiration behind this thread. A lot is said in the first posts. Feel free to read it.
 

Similar threads

Back
Top Bottom