the tag send data with FP_INDEX , this value in CLE is available between 700 and 900? why? In the CLE, may I change this value is between 740 and 760?thank you!
When receiving the signal the system sees the preamble and attempts to align itself to the incoming message. It then performs a correlation between the incoming signal and itâs initial guess. The results of that correlation are what you see in a CIR plot with the first path peak typically around 740-760.
The zero point for this process really doesnât matter much as long as the correlation peak is somewhere within the time period covered. It doesnât technically matter if itâs at sample 1 or sample 1000 as long as itâs included. Since where to put the start is based on an initial guess the logical choice would be to put the expected peak in the middle of the time period, this would allow for the largest possible error in that initial guess.
But we are looking at very weak signals. The chance that the initial guess is too early is very slim, the chances that itâs too late are far higher. Reflections are often stronger than the direct signal, if the initial guess ended up being based on a reflection then the true signal will be earlier. So rather than putting the initial guess in the middle of the time window it is instead put at a point roughly 3/4 of the way along.
It turns out that the initial guess is actually very accurate and so the FP_Index typically ends up being very close to 750.
1 Like
Hello, I have also been researching the CIR issue recently. Although I understand that the DW1000 aligns the CIRâs first path, why does the CIR always end around 850ns? What is the reason for this? I am using Channel 2, PRF=64, Preamble=1024, and Bitrate=110Kbps.
It doesnât always end there. It depends on the environment, in a larger or more reflective environment the ânoiseâ in the CIR after the first path will continue for a lot longer.
Thank you for your response. I understand your point, but it seems that in LOS conditions, the duration is shorter. Is this mainly determined by the duration of the received signal? Or, if the environment is wider, will it also become shorter? But is there a threshold for how short it can be? In contrast, in NLOS conditions, the duration is longer due to multipath effects, which is undeniable.
The âlengthâ of the CIR does not depends on the length of the frame. On DW1000, the CIR is provided only based on the Preamble segment, the other data (SFD, PHR, Data) are not shown.
If you increase the preamble length, you will see that the âCIR lengthâ will not increase, but the accumulated power will increase (the amplitude of the CIR). The CIR data should be taken as it is: the estimation of the Channel Impulse Response, this is not really a temporal view of the signal.
Back to basis, from a signal processing point of view, CIR is the signal you receive when you send a 0-lenght time signal with an infinite power (aka Dirac function). In UWB, we leverage the fact that the signal sent can be seen as a train of small spikes, so a train of âmini Dirac-lookingâ spikes. So after some signal processing stuff, the chip can provide an estimation of the CIR, according to the several spikes it accumulated.
Thank you for your answer, I understand