I'm spending some time in understanding L1 PHY development and have some questions on the start of packet detection in an IQ capture

Let's say I have a WiFi IQ capture which consists of 600 packets and want to extract indexes of start and End of each WiFi Packet.

When I plot abs(IQ_samples), it is like

WiFiPacket|NULL|WiFiPacket|NULL|NULL|NULL|WiFiPacket|NULL|WiFiPacket ....,

so what I have done is demodulator which gives me the plot ranges between 1 and -1, Since the energy of NULL is very much close to zero compared to the max_value of abs(IQ_sample of each packet).

So my packets have syncwords and Im trying to find the start of my packet by doing cross corelation according to my syncword that every packet have, but the plot of my cross corelation isn't giving me correct plot, it should gives me a high value when there's a syncword found in my demulated IQ samples that data ranges between 1 and -1. My main confusion is why the cross corelation isn't giving me a proper plot to see when my packets starts? I mean proper plot which represents a high value when there's a syncword found in my IQ demodulated samples, else would give me a low value ..

any help please how could I detect the first and the end of my wifi packets? thanks alot.

Im using SDR sampler for sampling ..

  • $\begingroup$ @Marcus Müller could you help me please? $\endgroup$ – Ryan Jun 11 '20 at 19:03
  • $\begingroup$ @Saira any assistance would be appreciated $\endgroup$ – Ryan Jun 11 '20 at 19:03
  • $\begingroup$ Have you removed carrier offsets or do you need to determine that as well in your correlation? $\endgroup$ – Dan Boschen Jun 11 '20 at 21:31
  • $\begingroup$ @DanBoschen didn't understand your question? in my cross corelation I do correlation between the IQ demulated samples (it's a vector) with my syncword which it's constant vector like 11-1-1-111 , my syncword is constant for every packet and I already know it "it's fixed value that I know it" $\endgroup$ – Ryan Jun 12 '20 at 19:56
  • $\begingroup$ If you haven't determined what the carrier offset is the correlation result will be reduced and will be rotating (or non existent if the offset is large enough). Given the transmitter and receiver are not synchronized there will be a difference in the actual clock frequencies and estimate of what the RF carrier frequency is. This needs to be resolved so my question is if this was already taken care of for you? $\endgroup$ – Dan Boschen Jun 12 '20 at 20:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.