Prev: Problem 2.10 in "An Engineers's Guide to MATLAB" 2nd edition
Next: removing elements from N-dimensional array
From: National Key Lab of ISDN ? on 28 Dec 2009 21:36
In simulation an OFDM system over Fading Channel, I got a problem with the ideal compensation.
The parameters used to setting the fading channel is:
SymbolRate = 20*1.0e6;
Ts = 1/SymbolRate; % Sampling Time
Tau = [0.0 0.4 0.9]*1.0e-6; % tap delay in seconds in SUI-3
PdpdB = [0 -5 -10]; % power-delay profile: power in each tap in dB
Pdp = 10.^(PdpdB/10);
Fdop = 100; % Max Doppler Frequency

and the channel is generated as follows:
ch = rayleighchan(Ts, Fdop, Tau, PdpdB);
ch.ResetBeforeFiltering = 0;
delay = ch.ChannelFilterDelay;

assume that we can get the channel impulse response by the operation:
ChTap = Tau/Ts+1;
h = zeros(1,N);
where N is the Number of subcarriers in an OFDM symbols
h(ChTap) = ch.PathGains;
H = (fft(h)).';

at the receieve side divide the recieved signal after passing the fading channel and the awgn channel, and using the 1./H to compensate the fading. However the BER performance is not improved at high Eb/n0, it is near flat when the Eb/n0 is higher.

I can not find the problem with these code, I need you help, thank you very much.
 | 
Pages: 1
Prev: Problem 2.10 in "An Engineers's Guide to MATLAB" 2nd edition
Next: removing elements from N-dimensional array