homemade LDE

Hi! As we know, LDE algorithm scans ACC_MEM and detects FP. The ACC_MEM array has a timestep of 1ns, while LDE detector’s resolution is 64 times better, according to 0x15’th FP_INDEX register value:

The first path (or leading edge) is a sub-nanosecond quantity. Each tap in the accumulator

corresponds to a sample time, which is approximately 1 nanosecond (or 30 cm in terms of the
radio signal’s flight time through air). To report the position of the leading edge more
accurately than this 1 nanosecond step size, the index value consist of a whole part and a
fraction part. The 10 most significant bits of FP_INDEX represent the integer portion of the
number and the 6 least significant bits represent the fractional part.

So, where does the “hidden salt” reside? Can I calculate that resolution on my own, analyzing ACC_MEM?


Hi, I have the same doubts as you. Have you solved them?
Best regards

Well clearly it works, 1 ns would only give 30 cm accuracy and the part exceeds that. Although it certainly isn’t 64 times better than that, individual ranges are in the few cm range at best not +/- 5 mm.

As to how, don’t expect decawave to give you an answer, that’s their IP.
I’d guess it’s based on the gradient of the spike and possibly matching the accumulator pattern to the expected ideal response in order to estimate where the signal started.

Thanks a lot, dude, you are so nice!
By the way, do you know how to get the value of RX_RAWST?
Maybe we can compare the value of RX_RAWST and RX_STAMP and find some rules?
Best regards