Iterative Channel Estimation for Multi-User OTFS Uplink Systems With Superimposed Full Pilots

Existing pilots based channel estimation schemes for orthogonal time frequency space (OTFS) systems usually require sufficient guard intervals to avoid pilot-data interference, which inevitably reduces the spectral efficiency (SE). This SE loss can be remedied by using superimposed pilots (SP). In this paper, we propose an effective superimposed full pilots (SFP)-aided iterative channel estimation (SFPCE) scheme for multi-user OTFS uplink systems. First, we derive a multi-user minimum root mean square error (MMSE) channel estimation algorithm and a message passing (MP) data detection algorithm based on the proposed multi-user SFP structure. Then, a de-interference iterative detection (DIID) algorithm is proposed to improve the accuracy of channel estimation and symbol detection. The DIID algorithm only requires a few iterations to achieve optimal performance. Besides, a hybrid superimposed pilot (HSP) structure is introduced to estimate the delay/Doppler taps without sacrificing the SE. Simulation results demonstrate that the accuracy of channel estimation and symbol detection can be significantly improved after several iterations. Moreover, the proposed SFPCE scheme is superior to the superimposed single pilot-aided channel estimation (SSIPCE) and the superimposed sparse pilot-aided channel estimation (SSPPCE) schemes in the bit error rate (BER) performance.