Low complexity linear equalizers with maximum multipath diversity for zero-padded transmissions

In wireless communications the fading multipath channel attenuates and distorts the transmitted signal. This makes equalization of the frequency selective channel of utmost importance. To exploit the full diversity provided by the multipath channel, maximum likelihood (ML) decoding is usually employed, which is computationally complex. In this paper we show that a specific linear zero-forcing equalizer is capable of benefitting from maximum multipath diversity in systems where the transmitted blocks are separated by zero guard intervals of length greater than the channel length (the well-known trailing zeros approach (Wang et al. (2000))). Furthermore, we exploit the banded Toeplitz structure of the channel matrix to reduce the complexity of the equalization process, and quantify the reduction in complexity. Simulations corroborate our results.