Hi Markus,

Well, the shortest possible pulse has duration 1/sampling rate and can
be sent with 

0…010…0

Now, that pulse has incredibly little power; and thus, the LO leakage
and all negative effects will probably make that unnecessary hard to
detect. I really recommend pulse compression:

Can't you just generate a random vector of +1/-1, use it as transmit
signal (embedded in a few 0s to allow the USRP to ramp up & flush out
remaining samples in the DSP chain)? In the receiver, you'd want to
correlate against that sequence – and luckily, correlation and
convolution are the same but for direction (and complex conjugate, but
1 and -1 are pretty real, so, nothing to conjugate), so just use a FIR
filter with the same sequence, just reversed().

You'd get the same resolution (because the cross-correlation would
reduce to a single sample, too, as long as your random sequence isn't
periodic), but you get n times the power, hence n times the SNR, n
being the length of the sequence.

Best regards,
Marcus

PS: Have you played around with gr-radar?
https://archive.fosdem.org/2018/schedule/event/radar/
On Fri, 2018-08-10 at 14:57 +0200, Markus Heller wrote:
> Dear list,
> 
> I wonder how I could generate one single shortest possible pulse on my
> USRP2 with GNURadio. 
> 
> Normally all the blocks would always generate an entire sequence, a
> continuous signal. 
> 
> I want to generate (and analyze) the shortest possible signal. 
> 
> Any ideas? 
> 
> The background: I want to generate a specific signal and record it's
> reflections. Some very basic radar experiments. 
> 
> Thanks!
> 
> BR / vy73
> Markus
> dl8rds
> 
> _______________________________________________
> Discuss-gnuradio mailing list
> Discuss-gnuradio@gnu.org
> https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

Attachment: smime.p7s
Description: S/MIME cryptographic signature

_______________________________________________
Discuss-gnuradio mailing list
Discuss-gnuradio@gnu.org
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

Reply via email to