Subject:
|
Re: Nic_1
|
Newsgroups:
|
lugnet.robotics
|
Date:
|
Fri, 10 Feb 2006 20:35:40 GMT
|
Viewed:
|
1945 times
|
| |
|
|
In lugnet.robotics, Matthew Ruschmann wrote:
> In lugnet.robotics, Claude Baumann wrote:
> > If you want to see Nic_1, a sound following robot, go to
> > http://www.convict.lu/Jeunes/ultimate_stuff/Nic_1.htm
> >
> > Nic_1 runs a program on the RCX that first samples two RCX sensor input ports at
> > 36000 Hz (there is no mistake here !) to receive 300 samples from both
> > microphone channels. Then the program operates a fast and simplified
> > cross-correlation, yielding the phase-lag between both audio-signals. This value
> > is used to indicate the direction towards which Nic_1 is supposed to turn. The
> > program prunes erroneous values that might appear due to room reverberation. The
> > program was written with Ultimate ROBOLAB.
> >
> > The video shows Nic_1 whilst tracking radio-music. It also could be a music
> > instrument or a human voice.
> >
> > Enjoy.
>
> Hey that's pretty amazing for the standard RCX firmware. I guess calculation
> time doesn't really matter though.
Calculation time is very short -regarded the H8 limits- On the video and the
program diagram you can conclude that the robot turns about every 2 seconds.
Deduce the 0.8 sec pause and you have something around 1.2 seconds for a
cross-correlation of 300 samples.
|
|
Message is in Reply To:
 | | Re: Nic_1
|
| (...) Hey that's pretty amazing for the standard RCX firmware. I guess calculation time doesn't really matter though. I'm a graduate EE taking estimation theory this semester so I'll keep this close to my heart. :-p I can't wait to try 2 mics on the (...) (19 years ago, 10-Feb-06, to lugnet.robotics)
|
5 Messages in This Thread:
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
