Channel Estimation in OTFS Systems by Leveraging Differential Modulation

By introducing orthogonal time frequency space (OTFS) modulation, the challenge of channel estimation (CE) in high-speed scenarios is alleviated. Nonetheless, the existing embedded pilot-based CE in OTFS systems significantly consumes user equipment (UE) energy and reduces system spectral efficiency (SE). Meanwhile, the superimposed pilot-based approach encounters the substantial challenge of superimposed interference. To tackle these issues, inspired by the advantages of differential modulation, differential modulation-aided CE in OTFS systems is proposed in this paper. At the UE, transmitted data is modulated using differential modulation to eliminate the need for pilot transmission and guard interval insertion. This strategy effectively reduces UE energy consumption and improves system SE. Especially, differential modulation is applied in the delay-time (DT) domain, which is less susceptible to Doppler spread compared to the delay-Doppler (DD) domain. At the base station (BS), employing a decision feedback-based scheme based on differential demodulation achieves the initial features of CE. Then, by leveraging the significant features of wireless channels in the DD domain, we design a lightweight network to enhance CE accuracy. This lightweight network consists of only one hidden layer, with the number of neurons in the hidden layer being identical to that of the input layer. This is attributed to the fusion learning involving differential detection-based decision feedback CE and the lightweight network assistance. Simulation results indicate that, in comparison to classic CE methods in OTFS systems, the proposed method enhances system SE, reduces UE energy consumption, and improves the CE accuracy.