Re: [simspark-devel] About Noise and Restricted Vision

[Sander v. D. <sgv...@gm...>]

Hey,

This noise model already causes more error for objects further away. With
noise of 1 degree on the angular components there is an error of 1.75 cm if
the object is 1 meter away, 17.5 when it is 10 meters away, et cetera. We
could also make the noise for the distance component dependent on the
distance, that may be more realistic (though I'm not familiar with the
performance of real vision systems in this aspect).

I am not sure about your last question. This has been implemented somewhere,
I'll have a look to see where and what it is (time I get into the code again
anyway :) )

Sander

On Sun, Jan 25, 2009 at 1:41 PM, Hedayat Vatankhah <hed...@ai...>wrote:

>  Hi,
> I think the errors should depend on distance too. Far distances should have
> more error than near distances. And I think the same thing was implemented
> lately in the spheres version(?!).
>
> Good luck,
> Hedayat
>
> *Sander van Dijk <sgv...@gm...> <sgv...@gm...>* wrote on
> ۰۹/۰۱/۲۳ 02:26:57:
>
> Perhaps we can just use the model of the old spheres simulation? Quoting
> from TEXT_INSTEAD_OF_MANUAL:
>
>   371 The noise parameters of the vision system are as follows:  372
>   373 - A small calibration error is added to the camera position. For each  374 axis, the error is uniformly distributed between -0.005m and  375 0.005m. The error is calculated once and remains constant during the
>   376 complete match.  377   378 - Dynamic noise normally distributed around 0.0  379   + distance error:  sigma = 0.0965
>   380   + angle error (x-y plane): sigma = 0.1225  381   + angle error (latitudal): sigma = 0.1480
>
>
>
> Sander
>
> On Fri, Jan 23, 2009 at 8:37 AM, Hedayat Vatankhah <hed...@ai...>wrote:
>
>>  Hi all,
>>
>> While TC will make final decisions until 13th of January, it seems that
>> they'll agree with introducing noise and also restricted vision for this
>> year.  If I remember correctly, restricted vision is implemented, so we can
>> use it right now (maybe it needs testing?!). But about noise, please help in
>> deciding about the noise model for each sensor and implementing it as soon
>> as possible, so that teams will have enough time to work on them.
>>
>>
>>   I think it would be nice if we release a new version of the simulator
>> before the end of February.  I think we should try to avoid introducing any
>> changes which could affect teams after that for this year.  (We might create
>> a separate brunch for maintaining the simulator for 2009 if we are
>> implementing new features in trunk.)
>>
>>
>>  So, what do you think about the noise model? Can we borrow from 2D? And
>> which perceptors should have noise for this year? All of them or just vision
>> or ...?!
>>
>>
>>  Cheers,
>>
>> Hedayat
>>
>>
>> ------------------------------------------------------------------------------
>> This SF.net email is sponsored by:
>> SourcForge Community
>> SourceForge wants to tell your story.
>> http://p.sf.net/sfu/sf-spreadtheword
>> _______________________________________________
>> Simspark Generic Physical MAS Simulator
>> simspark-devel mailing list
>> sim...@li...
>> https://lists.sourceforge.net/lists/listinfo/simspark-devel
>>
>>
>