Dear;

As you know, in the transmitter of OFDM system, to transmit the frequency components of the OFDM symbol, the signal must be represented in the time domain. This is accomplished by the IFFT. so you can not replace IFFT by FFT  theoretically and analytically.

The IFFT must  be used only if the information is already  transformed by the FFT. For example if y have a temporal signal, you may use the FFT to represent  it in the frequency domain. To return to time domain, you must use the IFFT because IFFT  is the inverse of FFT  transform.

I think of OFDM as a sort of cousin of the way FM stereo works (subcarrier for L-R signal and other subcarriers, e.g. subcarrier for background music services). You begin with a scheme arranged in a spectrum, with subcarriers distributed throughout a channel, in the frequency domain. Then you go through a transform, which converts this to the time domain. In FM, this is where the frequency modulation is introduced. And that's what you transmit. The spectrum of the FM signal is described by a Bessel function, which looks nothing like the neatly arranged spectrum before the frequency modulation step.

The IFFT in OFDM is needed to get all those thousands of subcarriers, arranged evenly across the channel, into the time domain, and out the transmitter and transmit antenna. So, I'm not sure how you'd do this the other way around?

Dear Researchers,

In the figure(1) of this attached Base paper, It is a 2 x 2 MIMO diagram, the output of one transmitter is processed by IFFT and other one is processed by FFT at the transmitter side.

And in the receiver side, one receiving antennas processes the data by using FFT and other one is processed by IFFT.

What I tried is instead of the parallel operation, I tried with both FFT at Transmitter side and IFFT at the receivers side. 

When I tried like this , actually BER ratio is decreasing . So I am asking this doubt. Please clarify my doubt. 

Dear,

Also in the base paper, The OFDM system is still employing IFFT at the transmitter and FFT at the receiver. The FFT that is employed at the transmitter (in the base paper) has a reversed operation for ICI cancellation purposes. This works when it is assuming that both 1st and 2nd branches are combined coherently without interfering with each other at the receiver side.

Dear

I think your question becomes: why we use IFFT at the transmitter and not FFT. Remember that signals need to be modulated by say N-QAM of the orthogonal subcarriers. Mathematically, the process can be represented by IFFT. In transmitters of OFDM, the OFDM symbol is constructed in the frequency domain by mapping the input bits on the I- and Q- components of the QAM symbols and then ordering them in a sequence with a specific length according to the number of subcarriers in the OFDM symbol. That is, by the mapping and ordering process, one constructs the frequency components of the OFDM symbol. To transmit them, the signal must be represented in the time domain. This is accomplished by the inverse fast Fourier transform IFFT.

You received good explanations, however if you had your initial data settings in time domain instead of frequency domain which is largely used in simulation you are supposed to apply FFT at the transmitter then IFFT at the receiver and the BER/SER of the system will still be the same, However this case is only used in Analogy signal transmission.

It certainly appears theoretically wrong, because when we consider each IFFT block sequence, then that is just a single sequence. In order to tie each of the IFFT complex time domain block sequences to their frequency (sub-carrier) components in frequency domain, it is expected that each time domain IFFT block sequence would need to be 'periodic' (repetitive/repeating). And it seems that the various details about the IFFT OFDM technique does not teach anything about repeatedly transmitting the same IFFT sequence (to have it periodic). That is, sending of each IFFT block sequence just once probably does not pass as 'OFDM', because an OFDM symbol associated with the frequency domain is expected to be tied to a 'periodic' form of the IFFT block signal in the time domain.