How to use RLS with complex numbers as the input?

[engomar]

I import a file which have complex numbers, and I want to filter it using complex RLS adaptive filter, but it always freeze at the filtering and update then it returns one value in the matrix and all the other numbers are NaN, so how can I fix this.

The second thing is when I use FBLMS, it always return error 42020 unless I specify the filter length= number of the input which is huge number, so how do I specify normal filter length.

[BobRodge]

Hi engomar,

There is a detailed description of how to choose filter length in the Adaptive Filter Tollkit Help file. 

http://zone.ni.com/reference/en-XX/help/372357B-01/lvaftconcepts/aft_choose_filterlength/

This page has several links to other important factors of what filter to choose. Under Computational Resource Requirements, it lists the memory usage for different filter types. RLS is the most memory intensive, which could be causing the freeze. Maybe try a LMS to see if this helps.

[trusted47]

hello engomar,

i have designed the RLS algorithm in labview as my final year project. i havent used any toolkit, i have implemented the formulas for tap weights, gain, phi matrix and error by utilizing basic linear algebra. the problem im having is with the desired input ( d(n) ).  i have applied sine + Gaussian white noise to the inputs but there is a problem. refer to this link. if you could please help me out. I've been stuck here for a while now! if you need the code tell me!

thankyou!