Hi
I would like to implement MUSIC DOA and compare the result to the SDK's method ( 3D FFT with zero padding ), I use AWR1642 with 2T4R chirp to implement 8 virtual antenna ( same as the SDK demo ). My question is:
How to get or make the s(t) of the 8 element ( virtual antenna ) for a selected range/velocity? Currently I use the Azimuth static heatmap result and choose a right index of range to get the 8 complex number ( AMP and PHASE ), than I calculate the range index frequency and use the 8 complex number to generate 8 time domain signal s(t) and than use Matlab to get the MUSIC spectrum. If the target is only one in that range with zero speed, I can get a very obvious peak in MUSIC spectrum in the right angle, but if there are more than one target for example 2 target, the result is not good enough and not stable and worse than the 3D FFT method even though the target is separate larger than the angle resolution. It is very strange. Does the method that I make the s(t) is correct or not if I want to MUSIC to calculate the DOA in a fix range and speed? or you have any suggestion to implement MUSIC?
