Interference cancellation for space-frequency OFDM MIMO systems: iterative decoding

One major assumption in most existing space-frequency (SF) OFDM systems is that neighboring OFDM sub-channels (SC) have the same channel transfer functions (CTF). However, this assumption might be not feasible for small number of SC or long code block length because in such cases the channel gains between adjacent SC will not be approximately constant. Even if the number of SC is large enough, in some parts of the OFDM spectrum, or in a severe multipath environment, the variation of the channel gain cannot be ignored. Because of this the linear ML decoder in Tarokh et al. (1999) cannot be readily used to achieve maximum diversity, and it will cause an irreducible error floor in the high signal-to-noise ratio (SNP) region. To this end, this paper proposes an interference cancellation (IC) signal detector to reduce the error floor. While the computational complexity of the IC detector is a little higher than that of the conventional ML linear detector, the new IC detector provides much better performance by subtracting the interference resulting from variations between the adjacent SC.