[simpits-tech] Motion...

[Joseph Fagner]

Good post Gordan, give us a lot to ponder.  I knew motion was much more
complicated than the simple idea I mentioned, but it has stimulated a lot of
good posting.  Seems to me your last point is the most poignant of all:  No
motion may be the best motion.

Jay

-----Original Message-----
From: simpits-tech-admin@simpits.org [mailto:simpits-tech-admin@simpits.org]
On Behalf Of Gordan Sikic
Sent: Monday, January 20, 2003 4:13 AM
To: simpits-tech@simpits.org
Subject: Re: [simpits-tech] Motion...

   Hi,

 >Think about this for a minute (granted I realize this is probably overly
 >simplistic and the math is infinitely more complex):  Most software out
 >there has the 4 axis of data (not to be confused with the axis of evil)
 >one might need for motion.  Couldn't one just write some software to
 >read this data and translate that to motion?  The data to which I refer
 >is (1) pitch and (2) roll  (X and Y axis); via your artificial horizon
 >instrument.  (3) Yaw (Z axis), via your compass and (4) altitude via the
 >altimeter.
 >
 >

You are right, it's not that simple.

You need angular rates as well as accelerations. Making motion platform
based on  attitude of the AC would be waste, believe me. Generally
speaking, with the motion platform one hopes to imitate (inertial)
accelerations that pilot feels. Simple example acceleration on the
runway during take-off. During that period pilot feels sustained
acceleration on his back. The sense of acceleration is imitated by
rotating the platform slightly backwards, so the weight of the  pilot
generates the force on the back. That rotational shift is sudden, and
later the platform is slowly returned to middle position. Thus,
imitation of just one linear acceleration is achieved completely
different type of motion: sudden rotation, followed by slow return to
neutral position. BTW, this is the basic type of operation for
calculation of required motion platform movement.

In brief, you need output of the state vector from the simulation
program in order to calculate required motion of the platform.

Other than this, you should think of something completely different:
Motion platforms are _very_ power hungry, you might easily end up with
facility which draws couple of  tens of kW of power.
Yet another problem is safety. Think of the following: it is very easy
to start the fire in the kitchen with electrical oven of just couple kW
(or even less), and here you are dealing with much more powerful equipment.

Yet yet additional problem are foundations, because something must
support complete construction, and take over complete forces and moments
generated by the moving platform.

There are many problems, but this doesn't mean that there is no solution :)
If you really want to make motion platform, do read appropriate chapter
in "flight simulation", book that is mentioned few days before. It is
excellent reading, but be warned that it does not provide us with
solutions, but presents the problems in very concise way, and just
guides towards the solution. But in any way it is a "must have" book.
I've been reading it from the cover to cover and all over again many
times since 1995 (when I bought it :)

Also check web for motion platforms where you will find some standards
  you should achieve. one url is www.fokercntroll.com, and the other is
(if I recall correctly, I'm off line now) www.mug.com.

Personally I'll take following approach:
Generally, visual system provides low frequency sensations, and use some
kind of rocking chair to generate high frequency sensations (like
turbulence, or vibrations, ...). It is also possible co combine it with
some form of g-suit. This is not as good as well made motion system
coupled with nice visualization, but it is _much_ better than wrongly
made motion platform, and in any case  infinitely safer.

And at  the end, one true story:
(I don't know the details, what kind of  AC, how many degrees of freedom
in motion platform, nor how many pilots, ... :)
There was a project of evaluating different algorithms for driving
motion platform, and within the project many (professional) pilots were
flying on the simulatior, every flight with different algorithm applied.
Needles to say that pilots didn't know which algorithm was to be applied
for any of the flights. After every try, they were asked to describe in
words how they felt, and to place a numerical mark. After the data were
analyzed, it turned out that many pilots (more than few, anyway) said
that the best results were achieved with algorithm "number 5". The funny
thing was that algorithm in question corresponded to the case where the
motion platform didn't move at all!  (of course, pilots didn't know that
fact while "flying")
This fact presented the problem also. Far to many pilots claimed that
nonmoving motion algorithm was best, so it had to be explained. The
explanation was as follows: the visualization was made very well, the
screen was wide, and clues from visualization contained enough data for
the brain to construct complete motion. When clues from the motion of
the platform were added, the brain was confused by the data provided by
the  platform motion, since it was not able to "correlate" it fully with
data from the visualization. The result: bad sensation by the pilots :)

I know this was just a bunch of unconnected thoughts but I do hope that
it helped a bit...

ciao,

Gordan


_______________________________________________
Simpits-tech mailing list
Simpits-tech@simpits.org
http://www.simpits.org/mailman/listinfo/simpits-tech
To unsubscribe, please see the instructions at the bottom of the above page.
Thanks!