Channel Estimation for mmWave High-Mobility Systems with 5G New Radio OFDM

Time-varying channels are significantly influenced by the Doppler effect, which leads to rapid changes in channel gain and requires Doppler frequency estimation to compensate for channel phase shifts and improve communication quality. In this paper, we propose a novel fifth-generation (5G) new radio (NR) orthogonal frequency division multiplexing (OFDM)-based transmission structure for time-varying channel estimation in high-mobility scenarios. By designing an appropriate subcarrier spacing and slot format, we ensure that the pilot signals remain nearly invariant within a single slot and exhibit rotational invariance between different slots. Leveraging this rotational invariance, we introduce a novel algorithm based on Vandermonde-structured tensor decomposition, which is non-iterative and has lower computational complexity and higher robustness than other tensor-based algorithms. Moreover, we provide a theoretical analysis of the uniqueness condition of tensor decomposition, proving that the proposed algorithm has strong feasibility and requires low pilot overhead. We also analyze the mean square errors (MSEs) of the parameter estimates and present a concise derivation of the Cramér-Rao Bound (CRB). The results demonstrate that the proposed algorithm significantly outperforms compressed sensing (CS)-based methods and other tensor-based methods in terms of parameter estimation performance at medium to high SNR. Furthermore, the proposed algorithm, based on the instantaneous channel model, offers higher channel estimation accuracy than the Kalman filtering-based algorithm, which relies on statistical channel models. Simulation results with channel data generated by Wireless InSite, which constructs a real-world scattering environment, demonstrate the high estimation accuracy of the proposed algorithm, validating its effectiveness in practical scenarios.