[Flii] New version of FWiidom in svn

Mon Apr 2 18:49:03 EDT 2007

yes, i was in a bit late today :)

reading the wikipedia article it becomes clearer how acceleration can be
estimated from the three accelerometer (i'll stop calling them gyroscopes
now) outputs. One interesting point i noticed was that as the outputs from
the accelerometers increase (in the positive or negative direction) their
accuracy decreases. So while it may appear that the z-axis is redundant,
when either the x or y axes are at an angle close to 90 degrees from the
vertical, it would be more accurate to include the z axis data in
calculations as it would be closer to 0 degrees from the vertical.

I also didn't realise the accelerometers could output a normalised reading
above 1g, so with this in mind it should be possible to calculate
acceleration independent of gravity effects, if we assume that the
acceleration occurs in a 'jerk' and the rotation can be re-calibrated at the
end (which shouldn't be too delayed considering the finite space people have
to accelerate things given the length of their arms).

i don't have any more time to look at this now (i think i probably need a
good nights sleep...) but i'll definitely give it some thought over the next
few days and let you know what i come up with...

also, the whitepaper on gesture recognition is completely useless for anyone
willing to understand the mechanism, but i'm sure there's more useful papers
out there.

Rob

On 4/2/07, Adam Robertson <oldskool73 at gmail.com> wrote:
>
> err, yeah, all three are in there, I may not be passing them all through
> to flash though as i assumed it wasn't all needed :)
>
> A
>
>
> On 4/2/07, Pete Hobson <pete at freesome.com> wrote:
> >
> > hey adam,
> >
> > can i just confirm you have access to *both* the live acceleration data
> > , and the 1G *and* zero point calibration data.  All three are needed to
> > calculate rotation correctly  believe.
> >
> > On 2 Apr 2007, at 10:20, Adam Robertson wrote:
> >
> > wow, bet someones a bit tired today ;)
> >
> > so, I take it it worked for you, that's a good first step!
> >
> > this page my be helpful (or may be what you already know)...
> >
> > http://www.wiili.org/index.php/Motion_analysis
> >
> > ...in fact that page seems to have been updated since I last looked and
> > have the pitch/roll equations, which I think are the same as we are using?
> >
> > A
> >
> >
> > On 4/2/07, Rob Bateman <rob.bateman at gmail.com> wrote:
> > >
> > > Hey guys
> > >
> > > i've had a look at the .as files and the output you get from the
> > > conduit, and looked at how it reacts to the motion of the wiimote
> > >
> > > very interesting stuff, and after much waving about, i've made some
> > > observations which i hope will help with the development
> > >
> > > i've not had a chance to compile my own tests, (no as3 on my pc at the
> > > mo) but from what i've seen of your own, there seems to be some confusion
> > > over the three gyroscopic outputs (measured as gX, gY and gZ i believe, as
> > > output in your graph demo)
> > >
> > > each one appears to be outputting the sin of the angle of the
> > > normalised vector from the horizontal. (a normalised vector to a plane
> > > represents a vector of length 1 extending perpendicularly out of the plane)
> > > to elaborate:
> > >
> > > gX is outputting the sin of the angle of the normalised vector  of the
> > > right side of the wiimote from the horizontal
> > > gY is outputting the sin of the angle of the normalised vector  of the
> > > front of the wiimote (where the ir sensor is) from the horizontal
> > > gZ is outputting the sin of the angle of the normalised vector  of the
> > > bottom of the wiimote from the horizontal
> > >
> > > in each case, a positive angle is represented as being above the
> > > horizontal plane, and a negative angle below. So for example, when gY is +1,
> > > the angle is +90 degrees and the right side of the wiimote is pointing
> > > straight up.
> > >
> > > for the fields in your wiimotecontroller class, pitch can be
> > > calculated in radians from the arcsin of the gY value, and roll can be
> > > calulated in radians from the arcsin of the gX value. The gZ value is
> > > slightly redundant as it doesn't add any information to that already given
> > > by gX and gY, but i suppose it could be used to smooth out erroneous
> > > readings from gX and gY.
> > >
> > > From examining the fields in the wiimotecontroller class, i'm a little
> > > confused by the presence of relative acceleration values. I'm unsure how
> > > these could be calculated to sufficient accuracy with the information
> > > available, because of the redundance with orientation thanks to Newtons
> > > second law. The force (acceleration) being applied to the wiimote could not
> > > be distinguished by the gyroscopes from the force being applied by the
> > > earths gravity, hence the two vectors would simple multiply together and
> > > would appear as if the wiimote had been tilted. It may be possible to
> > > produce something that could be subtracted from the orientation readings
> > > given by the ir sensor, but only if the wiimote was pointing at the sensor
> > > bar at the time.
> > >
> > > I plan to get my pc set up so that in future i can run tests myself,
> > > to get a better picture of what is it possible to determine from the output
> > > available. In the meantime, i hope the above information is useful! :)
> > >
> > > cheers
> > >
> > > Rob
> > >
> > >
> > > On 4/1/07, Pete Hobson < pete at freesome.com> wrote:
> > > >
> > > > Nice one adam.
> > > > I'll have a look at this under parallels on the mac, will be
> > > > interesting to see if it works there.
> > > >
> > > > One quick question - does using c# bring in dependancies on the .NET
> > > > framework also, or can you use c# for non .net console apps also??
> > > >
> > > >
> > > > On 1 Apr 2007, at 21:41, Adam Robertson wrote:
> > > >
> > > > Hi,
> > > >
> > > > I've put my work so far on the new FWiidom c# wrapper in my svn @.
> > > >
> > > > http://www.fwiidom.org/svn/
> > > >
> > > > there's a quick readme and the start of some bare bones
> > > > documentation in there too, if you are intrested in beta testing it feel
> > > > free to have a play and let me know how you get on.
> > > >
> > > > Adam
> > > >
> > > > _______________________________________________
> > > > Flii mailing list
> > > > Flii at osflash.org
> > > > http://osflash.org/mailman/listinfo/flii_osflash.org
> > > >
> > > >
> > > >
> > > > _______________________________________________
> > > > Flii mailing list
> > > > Flii at osflash.org
> > > > http://osflash.org/mailman/listinfo/flii_osflash.org
> > > >
> > > >
> > >
> > >
> > > --
> > > Rob Bateman
> > > Senior Creative Developer
> > >
> > > Work: 0207 7804841
> > > Mobile: 07714 329 073
> > >
> > > rob.bateman at akqa.com     |     http://www.akqa.com
> > >
> > > AKQA, 1 St John's Lane, London EC1M 4BL, UK.
> > > _______________________________________________
> > > Flii mailing list
> > > Flii at osflash.org
> > > http://osflash.org/mailman/listinfo/flii_osflash.org
> > >
> > >
> > _______________________________________________
> > Flii mailing list
> > Flii at osflash.org
> > http://osflash.org/mailman/listinfo/flii_osflash.org
> >
> >
> >
> > _______________________________________________
> > Flii mailing list
> > Flii at osflash.org
> > http://osflash.org/mailman/listinfo/flii_osflash.org
> >
> >
>
> _______________________________________________
> Flii mailing list
> Flii at osflash.org
> http://osflash.org/mailman/listinfo/flii_osflash.org
>
>

-- 
Rob Bateman
Senior Creative Developer

Work: 0207 7804841
Mobile: 07714 329 073

rob.bateman at akqa.com     |    http://www.akqa.com

AKQA, 1 St John's Lane, London EC1M 4BL, UK.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://osflash.org/pipermail/flii_osflash.org/attachments/20070402/f8828849/attachment.htm