On Mon, April 14, 2008 6:16 am, Andy Guest said:
> Hey folks,
> I'm looking at adding some functionality to a player/stage simulation to
> simulate the transmission of sound between robots. I'm not sure if I
> be asking this on the users list or the developers list so please forgive
> if I've got this in the wrong place.
> I am running an experiment that involves robots making sounds, listening
> to other robots and mimicking what they hear. I'm not interested (for the
> time being at least) in having the simulation make sounds I can hear, I am
> just looking at simulating the transmission of sounds between the robots.
> I've got a bit of code written that simulates packets of sound and I'm
> trying to work out how to build it in to p/s to see how it works there. My
> code works in two distinct parts, effectively mirroring the breakdown
> between the client code running on the robot and the world simulation
> running on stage. My options are to set up an external system that the
> client code for the robots can talk to or to build my code into
> itself. Building it into player/stage would mean I could use all the
> threading code already built in which would be great but I can't work out
> how to do this. Truth be told I can't work out where to start, I can't
> anything useful in the documentation and I'm running round in circles
> looking at the code. Any clues on where to start ?
> (In case it helps my sound code works as follows - the 'client' part of
> the code tells the server what sounds are made at which location and asks
> the server what can be heard at a given location at the current time. The
> server maintains a list of sounds.)
> Andy Guest
No real comments on P/S specifics, but I just did this for our Python
Simulator, albeit in a very hacky fashion:
If I were to do it right, I'd give each robot a sound-source-device and a
sound-detect-device. The source would register its sound with a server,
and the detector would make a request to get what it can hear.
The Python code can generate the sounds in real audio, too (turn it on/off
via the gui menu). Also, we replicated the cartoon speech bubble, so you
can alternatively see what they are saying (and hearing).
For these experiments replicating (Marocco and Nolfi, 2006), it appears
that the fact that this simulated sound has a finite carrying distance may
For additional experiments (in P/S), it would be cool if robots would make
a bit of "noise" as they moved. One might see the evolution of language
emerging from the usefulness of detection of sounds in general.
Douglas S. Blank
Associate Professor, Bryn Mawr College
Office: 610 526 6501