A Nonlinear Phase Compensation Method for FMCW Radar Based on VNCMD

The frequency-modulated continuous wave (FMCW) radar, employing a dechirp receiver structure, is widely favored for short-range imaging and ranging due to its simplicity, cost-effectiveness, and high resolution. However, inherent nonlinear errors degrade its imaging performance. Current nonlinear error compensation methods typically necessitate a high signal-to-noise ratio (SNR) in the reference beat signal and are unable to extract nonlinear errors in scenarios where multiple reference targets are close to each other, which limits their application in practical engineering. To address this, a software correction approach based on variational nonlinear chirp mode decomposition (VNCMD) is proposed in this article. Initially, VNCMD extracts the time-frequency (TF) signal of the reference beat, which is then modeled as a high-order polynomial phase signal (PPS) and a multisine phase signal (MPS). Subsequently, these models are optimized using matched demodulation. Finally, nonlinear phase correction is achieved using time resampling based on the estimated error parameters. Simulation and real data confirm the effectiveness and superiority of this method over existing algorithms under low SNR conditions, thereby enhancing radar imaging quality significantly.