Accurate extraction of multicomponent linear frequency modulation (LFM) signal parameters, such as onset frequency, linear modulation frequency, amplitude, and initial phase, is of great importance in the fields of ISAR, cognitive radio, electronic countermeasures, and star-ground communications. However, the task of accurately extracting the characteristic parameters of a signal is challenging when it has an extraordinarily large bandwidth as well as cross or neighboring components in the time-frequency domain. In this paper, we first review the main current methods used for multicomponent LFM signal decomposition and their challenges, and then propose a novel multi-parameter feature parameter extraction algorithm. The algorithmrealizes the direct and accurate extraction of thefeature parameters of multicomponent LFM signals at ultra-low sub-Nyquist sampling rate for the first time. Moreover, the algorithm is optimized for the computational complexity and anti-noise problems in practical applications, so that it has high accuracy, high efficiency and good noise robustness. We also compare the algorithm with innovative and existing methods, and the results show that the algorithm has excellent performance in feature parameter extraction accuracy, noise immunity and computational speed.

翻译：准确提取多分量线性调频（LFM）信号的起始频率、线性调频斜率、幅度和初始相位等参数，在逆合成孔径雷达、认知无线电、电子对抗和星地通信等领域具有重要意义。然而，当信号具有超大带宽且在时频域存在交叉或相邻分量时，准确提取信号特征参数的任务极具挑战性。本文首先回顾了当前用于多分量LFM信号分解的主要方法及其面临的挑战，随后提出了一种新颖的多参数特征提取算法。该算法首次实现了在超低亚奈奎斯特采样率下对多分量LFM信号特征参数的直接准确提取。此外，该算法针对实际应用中的计算复杂度和抗噪问题进行了优化，使其具有高精度、高效率及良好的噪声鲁棒性。我们还将该算法与创新及现有方法进行了比较，结果表明该算法在特征参数提取精度、抗噪性和计算速度方面均表现出优异性能。