Use multi-dongles to speedup scanner

Jiao Xianjun putaoshu at gmail.com
Fri Jan 17 01:38:45 UTC 2014


Hi,

As a "side product" of this thread: "How to get IQ samples from multiple
rtl-sdr dongles in a synchronized manner",A Matlab script is released.

http://www.mathworks.com/matlabcentral/fileexchange/45024-rtl-sdr-multi-dongles-based-flexible-spectrum-scanner
or
https://github.com/JiaoXianjun/multi-rtl-sdr-calibration

The script can use multiple dongles to do scanning, and get significant
speedup compared to single dongle. (You can check reported time cost by
running the script with different number of dongles)
Another script uses multiple dongles to scan the same band, and combine
those results together incoherently.

I admit that before calibrating multiple dongles, the significant is
limited.

Just for FUN.

README is also pasted here:

Multiple rtl-sdr(rtl_tcp) dongles based Matlab frequency scanner.
1. multi_rtl_sdr_split_scanner.m
Have multiple dongles run concurrently to speedup scanning wide band by
scanning different sub-band with different dongle.
2. multi_rtl_sdr_diversity_scanner.m
All dongles scan the same band, and then results from different dongles are
combined incoherently.
------------Usage------------

1. Assume that you have installed rtl-sdr
(http://sdr.osmocom.org/trac/wiki/rtl-sdr) and have those native utilities
run correctly already.

For example, you have multiple dongles, please run multiple rtl_tcp in
multiple shell respectively as:
rtl_tcp -p 1234 -d 0
rtl_tcp -p 1235 -d 1
rtl_tcp -p 1236 -d 3
...

2. Then run script multi_rtl_sdr_split_scanner.m or
multi_rtl_sdr_diversity_scanner.m in MATLAB.

ATTENTION! In some computer, each time before you run script, maybe you
need to terminate multiple rtl_tcp and re-launch them again.

Change following parameters in the script as you need:

num_dongle = 2;
start_freq = 935e6;
end_freq = 960e6;
freq_step = 0.05e6;
observe_time = 0.1;
gain = 0;
sample_rate = 2.048e6;
...

------------Algorithm------------
Use high sampling rate for each frequency point. Then Auto generated narrow
FIR is used to suppress noise and extract signal.
At last, power of signal at each frequency is estimated and spectrum of all
frequencies is generated.



BR


Jiao Xianjun
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osmocom.org/pipermail/osmocom-sdr/attachments/20140117/02a9448e/attachment.html>


More information about the osmocom-sdr mailing list