Channel Estimation and Equalization in OTFS Transmission System with Random Padding

The OFDM waveform based communication relies on an assumption that linear convolution in propagation channel can be replaced (modelled) by circular one. Such assumption is not valid in the presence of high Doppler shifts, for example in high-speed vehicle communication, like high-speed trains (HST). Recently the Orthogonal Time Frequency Space (OTFS) modulation has been introduced for data transmission in channels with strong Doppler effect. In contrary to OFDM, the OTFS does not use the circular convolution assumption. This makes channel estimation and equalization not straightforward. In this work this issue is deeper analyzed in the context of OTFS communication systems with random padding (RP-OTFS), recently introduced. Namely, an influence of the channel estimation procedure, proposed for the RP-OTFS, on the achievable transmission bit error rate (BER) is investigated for the linear minimum mean square error (LMMSE) equalizer and efficiency of different applied interpolation schemes is compared.

I. Introduction

Future communication standards should offer, among oth-ers, high-bitrate and robust digital data transmission in high-mobility scenarios [1]–[3]. In order to ensure this, modern trends in communication systems shows constant growth of carrier frequencies and bandwidths. For example maximum operating frequency for the LTE communication system is 3.8 GHz, while 5G can operate on 50 GHz and more. It is expected that beyond 5G systems will use an even higher frequency range. In some publications the frequency range is estimated as from upper part of millimeter waves to lower part of sub-millimeter waves.