Dear Hua,

On Thu, 10 Sep 2009 23:13:30 -0400
Hua vvang <[email protected]> wrote:

> Dear all:
> I am a grad student in US, I have had lots of fun with e-pucks, but recently
> as I am working on the communication part, I chose to start from the ircom,
> since it is a good basis to work on. Meanwhile, I found lots of confusions,
> here are part of my doubts:
> 
> 1. Is there any documentation for this ircom module?

there is no extended documentation, but all the basics are explained on 
www.e-puck.org
and the same information is put in the readme.txt of the libIrcom package.
If you want to dive into the code, we can try to help you here.

> 2. In term of communication, what's the relationship between sampling window
> size and the packet size? (3 times? in the synchronization example the
> sampling window is 32, while the packet size is 12),  anyone give any tips
> on it is appreciated?

No there is simply no relationship. Each time a bit is transferred, robots 
transmit oscillating
signals at some frequency (freq1 means bit 0, freq2 means bit 1). To decode 
efficiently the signal sent,
a receiving robot samples 32 times the signal for a single bit.

According to our tests, 32 is a very good choice, just do not change that, 
unless you want to be slower, 
or risk errors in communication.

in the examples, packet size is 8 bits + 2 bits of synchronization + 2 bits of 
CRC.
Practically you get only 8 bits of data. 
What is really going on inside the robot, is that the emitter will transmit (8 
+ 2 + 2) bits and the receiver will sample the signal (8+2+2) * 32 times to 
decode it.

So, if you want to transmit 16 bits instead of 8, go into ircom.h and change 
ircom_wordsize to 16, for instance.
But you can choose any value you like.


I hope this helps you understand a bit better what's going on there,
regards,
alexandre campo.



> 
> Thanks.
> 

_______________________________________________
E-puck-dev mailing list
[email protected]
https://mail.gna.org/listinfo/e-puck-dev

Reply via email to