Litcius/Paper detail

DNN-Based Fractional Doppler Channel Estimation for OTFS Modulation

Lin Guo, Peng Gu, Jun Zou, Guangzu Liu, Feng Shu

2023IEEE Transactions on Vehicular Technology20 citationsDOI

Abstract

In this paper, we proposed a deep neural network (DNN) based fractional Doppler channel estimation scheme for orthogonal time frequency space (OTFS) modulation in the air-to-ground communication scenario with high-dynamic Doppler. Based on the zero-padded OTFS structure, the traditional pilot pattern with guard symbols is adopted. The received pilots in the OTFS domain are used as the inputs of the network to estimate the channel parameters which are used in the MRC algorithm to demodulate the signal. In our proposed method, it can achieve the similar performance with 14 dB boost of pilot energy comparing with the ideal channel estimation case, while the conventional method requires 30 dB higher. Both the accuracy and generalization ability of the DNN network are validated.

Topics & Concepts

DemodulationChannel (broadcasting)Orthogonal frequency-division multiplexingElectronic engineeringDoppler effectModulation (music)AlgorithmArtificial neural networkPilot signalEngineeringComputer scienceArtificial intelligenceTelecommunicationsAcousticsPhysicsAstronomyPAPR reduction in OFDMAdvanced Photonic Communication SystemsOptical Network Technologies
DNN-Based Fractional Doppler Channel Estimation for OTFS Modulation | Litcius