DSPRelated.com
Forums

Q) OFDM and BER verification

Started by lsw00kor February 18, 2007
Hello. Could I have some help & discussion please?

I built an OFDM simulator with IEEE 802.20 channel model. 
But, I'm not sure if my BER results are correct or not.

Thus, I checked BER of OFDM  in AWGN  first. 
Then, I checked BER of OFDM in very very low bit rate. For example when a
channel shows rms delay spread of 0.37usec, I tested with Rb = 1E4 bps.
Then, I checked BER of OFDM with higher data rate.

Here is my question: With OFDM, can we realize gaussian performance ( for
example, 16-QAM in Gaussian channel performance)? My result is that
always, OFDM BER is far away from Gaussian BER. I'm not sure if I did the
right or not.

I'd highly appreciate any suggestion or discussion.

Thank you very much. 
>Hello. Could I have some help & discussion please? > >I built an OFDM simulator with IEEE 802.20 channel model. >But, I'm not sure if my BER results are correct or not. > >Thus, I checked BER of OFDM in AWGN first. >Then, I checked BER of OFDM in very very low bit rate. For example when
a
>channel shows rms delay spread of 0.37usec, I tested with Rb = 1E4 bps. >Then, I checked BER of OFDM with higher data rate. > >Here is my question: With OFDM, can we realize gaussian performance (
for
>example, 16-QAM in Gaussian channel performance)? My result is that >always, OFDM BER is far away from Gaussian BER. I'm not sure if I did
the
>right or not. > >I'd highly appreciate any suggestion or discussion. > >Thank you very much. >
Hi, let us define the following equation r(t)=s(t)*H(t)+w(t) r(t)=received OFDM signal s(t)=transmitted OFDM signal H(t)=channel impulse response w= additive white gaussian noise. when performance of OFDM is studied only in AWGN, the above equation becomes r(t)=s(t)+w(t) In this case the plot of SNR vs BER matches the theoritical BER given in the literature,irrespective of modulation(QPSK,16QAM) used in OFDM. when channel is assumed along with AWGN, the equation is r(t)=s(t)*H(t)+w(t) depending on the channel properties, we see some delay spread in the received signal(your case). In this case, there is no way that the plot of SNR vs BER matches the theoritical BER. This is because, the thoritical BER given in literature(John G Proakis) is only in the presence of AWGN. In fact BER is worse in this case when compared with BER in the presence of only AWGN. I hope you got my point. Aamer