Go Back  PPRuNe Forums > Flight Deck Forums > Tech Log
Reload this Page >

New AI Design

Tech Log The very best in practical technical discussion on the web

New AI Design

Old 10th Nov 2009, 09:19
  #21 (permalink)  
 
Join Date: Nov 2009
Location: UK
Posts: 2
Likes: 0
Received 0 Likes on 0 Posts
Capt Pitbull is quite correct; there are, unfortunately, some errors in the New Scientist article. But the article was not written by the originators of the development work––it was written by a journalist.

The word ‘bunt’ does not appear in any of the work associated with the project––or in the presentation given at the week long Royal Aeronautical Society CEAS International New Innovations Conference held in Manchester.

Were any of you there?

Clearly not.

Because I had assumed––obviously, completely incorrectly, that the negative comments being posted on this next-generation thinking––were as a result of reading––and carefully assessing––the RAeS published academic conference paper on the subject––and not on a postage-stamp sized illustration of part of an experiment, in a half-page magazine article with, as you quite rightly point out, some holes in it.

This work (which is continuing) was conducted as a seven year full-time research and iterative design development programme, in very close association with senior airline and military establishments (and the cockpit display is only a part of the new systems produced).

(By the way: there most certainly are pitch indications in the display––but the IPRS system only displays information as the pilot needs it––again, that is all clearly covered in the conference paper which I'm sure the RAeS will be happy to supply you with.)

One of the key outcomes of the research was this: Adrenaline is the killer.

Go and do some research on that yourself––and whist you’re at it, look at some of the internationally acclaimed work done on the issues relative to LoC and CFIT (much of it funded by NASA and the FAA) by Shappell and Weigmann; Chris Wickens, and Stanley Roscoe.

What you will come to is a very clear conclusion––No amount of training will overcome the automatic sympathetic nervous system (a relict of our evolution): Once adrenaline floods into the bloodstream, thinking closes-down. The pilot cannot read one mechanical aesthetic instrument––let alone two. The accident statistics prove that.

So, the outcome of the research was that to eliminate the effects of adrenaline, we need systems that don’t require thinking.

And whilst you are doing some research, check the E/GPWS stastics––accidents still occur where ground proximity systems annunciate. (And we have looked at and addressed the reasons for that.)

Get used to one fact: change, is inevitable.

It is no longer acceptable––or necessary––for an airliner full of innocent people to be lost in circumstances of loss of control or controlled flight into terrain.

Let me repeat that: It is no longer necessary for LoC or CFIT to happen.

Loss of Control and Controlled Flight Into Terrain can be stopped.

Although there are military applications for the work very briefly reported in New Scientist––the overriding objective––the thing that drives it, is (and remains) to stop control of airliners being lost in weather extremes––and to stop airliners flying into mountains or into the ground short of the runway on approach.
DeltaRomeoWhisky is offline  
Old 10th Nov 2009, 10:05
  #22 (permalink)  
 
Join Date: Sep 2008
Location: Scotland
Posts: 894
Received 6 Likes on 2 Posts
Would it be possible to provide a link to this paper for those of us without academic journal subscriptions?
Jwscud is online now  
Old 10th Nov 2009, 10:58
  #23 (permalink)  
 
Join Date: Aug 1999
Location: England
Posts: 1,050
Likes: 0
Received 0 Likes on 0 Posts
DRW,

Look, my main beef is with the NS article. Its a pretty shoddy piece of work. I don't agree with the emphasis of some of the things you said in the original article but lets just clear the air on a couple of points.

1. There are plenty of awfully designed systems on modern aircraft. I applaud any attempt to improve the equipment we use.

2.
It is no longer acceptable––or necessary––for an airliner full of innocent people to be lost in circumstances of loss of control or controlled flight into terrain.
It never has been acceptable. Any attempts to improve the situation deserve to be supported.

What you will come to is a very clear conclusion––No amount of training will overcome the automatic sympathetic nervous system (a relict of our evolution): Once adrenaline floods into the bloodstream, thinking closes-down.
With all due respect, no ammount of engineering effort will produce a system that is 100% fault proof, no UI design will provide a system that is incapable of being misinterpreted / ignored. With all these things you reach a point of diminishing returns.

So, yes, obviously ultimately anyone can be incapacitated by fear. However, there is no doubt that solid training and experience give you a much better margin before you reach that point.

Ultimately if the system requires any control inputs at all then if someone is panicking or paralysed a safe outcome can not be gaurunteed.


The pilot cannot read one mechanical aesthetic instrument––let alone two. The accident statistics prove that.
Really? What do statistics prove exactly? If a pilot does not succesfully interpret an instrument, how does that prove the design is at fault. When an aircraft is in a million pieces, we don't even know if he was looking at it.

The point is, some of us have had in depth rigorous loss of control training. Most haven't.

Also, you can quote all the academic names you want. Those of us at the coal face in the industry know what the biggest single problem is; weak original training, further atrophied by lack of practice, coupled with complacency at both the individual and the corporate level.

Yes, we would love better equipment, but its only part of the solution and its not a panacea.

Get used to one fact: change, is inevitable
Very used to that concept thankyou. I've seen plenty of change during my time in this profession, (including a good chunk as an avionics instructor) and I've always been happy to grasp the bull by the horns. Unfortunately, accountant driven management practices within the airline community view everything as an opportunity to cut costs. When you look at the industry holistically it is apparent that, regardless of what you and I both think, losing a few hulls world wide anually is deemed acceptable. If this were not the case, the accountants would not be chipping away at our layers of defence. I've experienced first hand the uphill struggle to justify training time - even when the training content is mandated!

Finally DRW, I have to ask this, and its not a willy waving question, but genuine interest, what's your own flying experience? I'd just like to know where you are coming from.

pb
Capt Pit Bull is offline  
Old 10th Nov 2009, 11:14
  #24 (permalink)  
 
Join Date: Aug 2009
Location: England
Age: 35
Posts: 84
Likes: 0
Received 0 Likes on 0 Posts
I think some of you are getting bogged down with your experience somewhat. If you're an experienced IR pilot then you're clearly going to like the old method much better.

From a PPL student perspective though, I do think that my view is more objective than some, and from first impressions the new method does seem more intuitive. The proof of the pudding is that if I inadvertantly entered IMC and found myself in a horrible attitude - I'd be comfortable in using the newer option wheras on the spur of the moment, using the older model I think I'd probably turn the wrong way, or at least it would take me a while to decipher the gauge.

In my opinion the newer model does not allow for attitude confusion but the old model definately does. It you're an expert at reading the old style then so-be-it, but for new pilots like myself I'd be very happy to use the new one.
TheOptimist is offline  
Old 10th Nov 2009, 11:30
  #25 (permalink)  
 
Join Date: Jul 2000
Location: London
Posts: 207
Likes: 0
Received 0 Likes on 0 Posts
Well, I admit that I havent delved into the full research, but really I am struggling to understand how a number moving around a blank screen is clearer than the modern generation of synthetic vision panels that are being introduced. Denti posted a very good example, and many others are out there.

I believe putting an accurate picture of the outside world, relative to you, whatever your present attitude, is the best method. And these new synthetic vision displays are amazing. I fly the PlaneView on the G550 and it is like the Falcon Easy displays as well, you really have a good SA.
smallfry is offline  
Old 10th Nov 2009, 11:33
  #26 (permalink)  
 
Join Date: Mar 2009
Location: Inside
Posts: 285
Likes: 0
Received 0 Likes on 0 Posts
change, is inevitable.
Then why pursue it?
One Outsider is offline  
Old 10th Nov 2009, 11:34
  #27 (permalink)  
 
Join Date: Aug 2009
Location: GPS L INVALID
Posts: 580
Likes: 0
Received 0 Likes on 0 Posts
I still don't understand how the whole attitude indicator discussion is in any way related to CFITs - misinterpreting the ADI is probably one of the smallest contributors to such an event while poor situational awareness (especially about the position of the aircraft) resulting from or combined with erronous navigation are the most immediate causes.
STBYRUD is offline  
Old 10th Nov 2009, 12:43
  #28 (permalink)  
 
Join Date: Feb 2005
Location: Correr es mi destino por no llevar papel
Posts: 1,422
Received 1 Like on 1 Post
Well, this research was...

Originally Posted by DeltaRomeoWhiskey
...conducted as a seven year full-time research and iterative design development programme, in very close association with senior airline and military establishments
...in which...

Originally Posted by DeltaRomeoWhiskey
The participants were all current fast-jet, military logistics and civilian commercial airline (Airbus/Boeing) pilots with a total experience of 100,078 flying hours––and an IR was a minimum requirement for inclusion.
...and then ...

Originally Posted by The New Scientist
A conventional display, which tilts the artificial horizon when the aircraft banks, was misinterpreted by the 92% of the pilots, leading to erroneous and potentially lethal choices.
Basically, it's sales pitch at its worst - in aviation, anyway.

It's the same type solution we adopted when people got killed by misreading the needle or the counter - navigation display was introduced an moving map with it. So do we crash nowadays when the map shifts or we return to good old needles, pointers, CDI and radar vectors? The statement "we need systems that don't require thinking" is quite emblematic of whole project. It's all very well when your synthetic vision conforms to what really is outside, what happens when it doesn't? How does your not-thinking pilot recognize that system is not working properly when there are no flags? And if he recognizes the problem, wil he still be proficient enough to keep the aircraft upright by referring to backup AI?

So dear mr. Wilson, until issues of map shifts and navigational accuracy are eliminated for good, please do not come anywhere near the cockpits of manned aircraft with your synthetic vision. It will create more problems than it apparently solves.
Clandestino is offline  
Old 10th Nov 2009, 14:03
  #29 (permalink)  
 
Join Date: Jul 2000
Location: London
Posts: 207
Likes: 0
Received 0 Likes on 0 Posts
We are talking about two different things here. The New Scientist article refers to a whole new concept of the display.
Whereas the synthetic systems we already have are extensions and improvements on the existing display.
I can assure you that whilst I have extremely good SA now with my PlaneView cockpit, I can still follow needles for ILS, VOR and NDB's. I do however, have a lot better picture and reliability than in my fathers generation steam driven cockpits of old. Map Shift and nav accuracy? pfff... I dont remember the good old boys flying the comets, VC10's 707's in modern airspace with RVSM, MNPS etc requirements... Right place right day was often all that was required... Laser gyros and GPS are a damn sight better than an VOR with water in it... (see the BA franchise BMED report going into Addis not so long ago...)

I am all for development and innovation, to improve our work environment, but some ideas need a lot more work than others.
smallfry is offline  

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service

Copyright © 2024 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.