We control the horizontal
Posted by aogTuesday, 17 November 2009 at 20:00 TrackBack Ping URL

New Scientist has an article about artificial horizon gauges in airplanes. The claim is that the current model is a bad design, likely to confuse pilots in stressful situations which is precisely when it’s most important to read it correctly. They suggest a very different model which is really only possible with modern displays1.

Anyway what I found interesting in the article was this —

Dewar Donnithorne-Tait, president of the Canadian Centre for Unmanned Vehicle Systems in Medicine Hat, Alberta, says it is more likely that the technology will be used by pilots flying uncrewed aircraft from a control centre on the ground. “Anything that reduces the confusion a remote pilot feels because he is not in the aircraft is a great idea. This idea would be even more useful in a UAV than in a manned plane.”

That is, we now have systems in regular use in which we can test human computer interactions in the real world with significantly less risk (at least to humans). I wonder if that may turn out to be a general rule, or whether direct physical presence will become passé first.


1 The current style can be built completely mechanically with no computing elements. The new style requires both computation and a graphical display.

Comments — Formatting by Textile
Hey Skipper Thursday, 19 November 2009 at 12:08

Sorry, this whole thing does not wash.

Back in the day, when I was an instructor pilot, one of the things we routinely did was to have the student close his eyes and look down. Then the instructor would maneuver the aircraft in ways to promote spatial disorientation (e.g., relatively rapid bank in one direction, then slowly unwinding the bank while allowing the nose to fall). After 15-45 seconds of this, the instructor says: you have the plane.

Whereupon the student has to confirm the unusual attitude, which includes integrating airspeed and vertical speed (because, after all, the evil IP could have tumbled the attitude indicator), then regain aircraft control.

It isn’t easy, primarily due to having to overcome conflicting vestibular cues. But once having figured it out, which is, after all, what pilot training is all about, the misinterpretation rate went to nearly zero. And that includes extremely unusual attitudes; every combination of pitch and bank you can think of.

Try it yourself. Assume that heading and VSI confirms the attitude indicator. The rule is to move the stick in the shortest direction between the bank angle pointer (red arrow fixed at aircraft vertical) to the sky pointer (black arrow indicating earth vertical). Any problems interpreting the display?

The modern attitude indicator is extremely intuitive, for exactly the opposite reason Mr. Wilson states: In the real world, the horizon is the fixed, stable element against which all movement is measured. In pilot world, the aircraft itself — not the horizon — is the fixed, stable element against which all other movement is measured.

UAV’s could well be a different matter, because the UAV pilot is always pulling 1G.

SD mishaps almost always involve transitioning from visual to instrument flying, for reasons having nothing to do with the design of the modern attitude indicators (ie., any built since the 1950s), but rather due to a late transition to instruments, or basic lack of competence (a la JFK Jr.), and consequent inability to reject erroneous vestibular sensations.

Post a comment