Measuring RSS during packet reception

GNU Radio
1

Hi again gnuradioers,

as always, I would like to ask for your opinion before starting to write
my
own block.
Basically, I would like to add a feature to my receiver to be able to
calculate average signal strength during signal reception. When I
receive a
valid packet I would like to have a RSS value. My question is what is
the
best way to accomplish this in GNURADIO? Using tags or messages, …? IS
there anything similar already provided so that I can take a look?

Thank you and best

2

Any suggestions on this topic, opinions?

Best
Nemanja

3

The only way that is on my mind is to detect burst and store it in a
file,
and then to send a message to the block that will read that file, and in
case of successfull demodulation and decoding to calcualte signal power,
but it looks pretty robust and non elegant.

Cheers,
Nemanja

4

This looks a bit silly cause I am speaking with myself. Anyway, maybe I
can
do something with probe_avg_mag_sqr, but maybe somebody can tell me how
can
I access that block’s values from another block?

Best,
Nemanja

5

Hi,

thank you Adeel for your answer. I implemented almost the same approach
yesterday. I bring signal from USRP through rms and log calculator
directly
to the block for sampling symbol value. On every symbol I read RSSI and
send that value forward. As soon as I detect premble i start calculating
RSSi for the packet based on number of symbols.
The problem that I have now is how to ensure that samples from
demodulator
chain and samples from RSSi calculator come in the same time to the
symbol
decision block. How can I tune this delay and is it deterministic?

Many thanks,
Nemanja

6

just to add this: In scope sink signals looks pretty alligned, but when
I
print RSSI value for every bit they don’t match to waveform, so I
suppose
(and hope) that comes from the way how scope sink works.

7

Nemanja,

I built this block sometime back. The ideas is to construct a custom
RSSI
Measurement block having input

1: complex IQ-samples after USRP-SRC or before “AGC”
2: byte-stream after “digital_correlate_access_code_bb”. Output byte
stream
has data in LSB (1st bit) while Flag in 2nd bit shows header detection)
3: Calculate system/filters group-delay between the complex-IQ samples
stream and output-of-correlate-access-code. This delay will depend on
the
a) No. of filters in system e.g. FLL-Filter, PFB-Timing-Synch etc
b) No. of filter Taps
3: Delay the complex IQ-samples stream by amount calculated in 3
4: In Custom RSSI-Measurement Block use following logic
if (2-bit of correlate-access-code-stream) is High:
Take average of N-items from IQ-samples stream

where “N” depends on
a) No of symbol in packet
b) Samples-per-Symbol

Hope this helps

-Adeel

8

Nemanja,

The problem that I have now is how to ensure that samples from
demodulator chain and samples from RSSi calculator come in the same time
to
the symbol decision block. How can I tune this delay and is it
deterministic?
Yes, the delay is fix and can be determined by adding delay of all
“Filters” in Rx-chain. Moreover IIRC, corrrelate-access-code itself adds
64bits delay that u need to cater for in ur custom block.

On every symbol I read RSSI and send that value forward. As soon as I
detect premble i start calculating RSSi for the packet based on number
of
symbols
I will suggest that donot calculate RSSI for every symbol instead
calculate
RSSI for complete packet

-Adeel

9

Hi Adeel, all,

thank you again for your reply.
The structure of my receiver is following:

               /--->  filter  ---> quadrature demod  ---->

threshold
detector —> clock synchronization —> correlate ac —> deframer
usrp source
/
/


/
—> rms —> 10 log
----------------------------------------------------------------/

Inside clock synchronization block stream is “broken” it might be that I
am
not able to determine where the actual burst started (maybe roughly) and
that’s why I want to send information about rssi together with every
symbol
value (i will think about idea to use 6 upper unused bits to put rssi
inside).
With current configuration it looks like the only delay I need to
calculate
is due to filter at the begining? (all other blocks should not introduce
any delay). What should I do in order to get the order of filter?

Best,
Nemanja

10

Hi all,

finally I added only delay of 64 samples due to correlate access code
and
it works good. My filter is of order 165 and with that additional delay
the
rssi is wrong. Why there is no delay due to the filter?

Thank you