[SOLVED] BER Result For MC CDMA (OFDM-CDMA)

hi everybody

i have a bad problem with my code ....

First of all i write a matlab code for OFDM-CDMA system. the following picture shows the structure of the system :



but unfortunally the results for BER not stable .
for doppler frequency = 0.0001


for doppler frequency = 3Hz


notes:

1) i use the built in matlab function rayleighchan for implement the effect of Rayleigh channel.
2) i choose the uplink case (as the system structure showed)
3) i generate a OFDM-CDMA signal for each user then each signal enter alone [/COLOR]to rayleighchan then i add the all signals together then adding AWGN and finally enter to receiver.
4) i calculate the BER for user number one only
5) i make iteration for the code ,, and i reach to 500 times without reach to stable result
6) for checking i run the code without any channel (nor rayleigh nor AWGN) i get 0 errors.
7) i use an itu recommendation for parameter of rayleigh channel which is channel A for indoor to outdoor case
(you can type in matlab help for this channel simply it is 4 path channel with delay equal to (0 110 190 410)nSec and path gain equal to (0 -10 -20 -23) dB)

Any help will be appreciated

thank you to everyone read, comment or help in this problem

Hi Ammaralwazzan,
you can have a look on my attachment regarding Rayleigh Fading Channels in Mobile Digital Communication Systems Part I: Characterization. Also try to find his book as well. The book has title: B. Sklar , Digital Communications: Fundamentals and Applications, Englewood Cliffs, NJ: Prentice Hall, 1988. I think you can find it free, otherwise please inform me

Regards,

hi blerman13

thank you for your fast reply and thank you again for your attachment file

i read it ,, it contain a great diagram for types of fading channel but my problem is in matlab code specificlly, not in theory

unless that the big difference in the results are true due to the types of the channel that i used.

see this for doppler frequency = 0.0001 Hz:



so if the following information are specify the condition of the rayleigh channel that i used:

ts = 1 / (16e6); % sampling time due to OFDM Part

fd = 3; % doppler frequency in Hz .

delay = [0 110 190 410]*10^(-9); % delay of each path in Sec

gain = [0 -10 -20 -23]; % Gain for each path in dB

so what we can name this channel? frequency selective fading? falt fading? time invarient? time Varient? and so on.....

unfortunately i try to find the book but i failed.

thank you very much,, Best regards

salam alikom Ammar

1-for the first picture, ok do 10 run each of 500 times (or blocks or iterations) as an example, then take them on average i.e. BER_mean= mean( all BER values in 500 vectors in all runs) , thus
semilogy (Eb/no, BER_mean)

2-for the second picture, run on more number of bits to get ur curve more smooth

that what I know

thanks

SA

thank you for every person who reply in this thread,

the solution was so simple .... and for me its a advice for all users of Matlab:

simply shut down the matlab and then run the program again and start your code and i have the same curve that i get it in first run.

so the only explanation is that the matlab go back to its initial conditions and effect with the same fading behaivor on system.

the second think is the value of doppler frequency !! when i use a doppler frequency with 3 Hz .. the same curve i get but with variation

in the curve (not smooth) even when i use a large number of iteration (repeat the code and then take the mean of BER)

but when i change my Doppler frequency value to = 0.0001 (near to 0) i got a smooth curve

for clarification the doppler frequency refer to the relative movement between transmitter and receiver , so for doppler frequency = 0.0001

it almost fixed motion between Tx and Rx.

thank you again for blerman13 and ahmeddeia

can you give me the matlab code for checking purpose