Doppler Frequency Estimation by Parameterized Time-Frequency Transform and Phase Compensation Technique

Doppler frequency can reflect the dynamic properties of the radar target and thus provides an effective approach to analyze the target signatures. In this paper, a novel method is proposed to accurately estimate Doppler frequencies of multi-component radar signals. The method employs a more general model, i.e., the redundant Fourier model to represent various Doppler frequency trajectories. In the first stage of the method, a multi-component signal is decomposed into some individual components using a de-chirping method. Then, the parameterized time-frequency transform (PTFT) with a matched Fourier kernel is designed to estimate frequencies of the obtained components. With the estimated frequencies, the complex envelopes of the components can then be estimated through a joint-least-squares method. We show that the phase information of the obtained envelopes can be used for error compensation in the frequency estimation. The proposed method integrating the PTFT with the phase compensation technique can significantly improve the accuracy of Doppler frequency estimation. Both simulated and experimental examples are provided, indicating the potential of the method in analyzing Doppler signals with overlapped components. The future study will focus on the real applications of the method for radar target detection and monitoring, such as human vital sign detection and condition monitoring of industrial equipments (e.g., wind turbines).