I'm designing a multi user Direct sequence spread spectrum in SIMULINK as my project, with BPSK modulation, and I want to implement the technique of multi - user interference rejection using LMS equalizers. Basically I am connecting multiple users at the transmittter side, and picking one of them as my desired signal, trying to retrieve it back on the rceiver side, and calculating the error rate. I have used an AWGN channel. Below is the rough order of the blocks I used.
Data>bpsk modulator >product with pn seq ( spreading ) > digital filter > added with other users > awgn chanel > lms linear equalizer > product with desired usr's pn sequence ( despreading ) > digital filter > bpsk demodulator > error rate calculator
Below are the specifications I used.
Data generator : Bernoulii Sequence Generator, time period Tb = 0.6
Spreading code : PN sequence generator, time period Tc= 0.04
All filters used all Low pass, FIR filters, Rectangular window, Normalized frequency cut off = 0.95
LMS Linear equalizer : 16 taps, reference tap is 4
I used differenct generating polynomials for the PN sequence generators in each of the different user to distinguish them, also the data generator, i changed its Probability of Zero each time to distinguish the data also.
The error rate refuses to go below 0.3 and the worst part is, with every increase in the user, the error rate decreases, which is wrong. I am unable to trace the error in the design. I have posted a picture of the model for better understanding.
To summarize it, I have the following doubts :
1) why does the error rate decrease as my no of users increase ?
2) The Lowpass FIR filter gives the least error at 0.95 cutoff - is this nearing to an ALL pass filter ?
3) Does the equalizer contribute to the error ? Do I need to add additional blocks to this ?