Partial data spreading, an efficient and inexpensive approach to combating frequency selectivity in coded OFDM transmission

Frequency diversity is a proved remedy against the detrimental impact of the frequency selectivity of radio channels. In the case of the channel access scheme OFDM, the obvious and known approach to frequency diversity would consist in commonly spreading the data elements to be transmitted over all OFDM subcarriers. Unfortunately, this approach of full data spreading would render optimum, that is Maximum Likelihood data despreading in the receiver prohibitively expensive so that only less complex suboptimum despreading schemes based on Zero Forcing or on still quite expensive iterative procedures remain. In order to overcome this situation, we propose in the present paper the scheme of partial data spreading. In this scheme the data elements to be transmitted are partitioned into subsets, and the data elements of each of these subsets are commonly spread only over a corresponding subset of the OFDM subcarriers. In this way, optimum data despreading becomes feasible. Our simulation results show that, as compared to un-spread transmission and depending on the characteristics of the radio channel, a significant reduction of the required SNR can be achieved. We consider OFDM transmission with rate 1/2 convolutional coding, QPSK data modulation, and soft input Viterbi decoding.