基于瞬时频率曲线拟合和局部搜索组合方法的三阶多项式相位信号参数估计

三阶多项式相位信号是一类非常重要而常见的非平稳信号。在非合情况下，针对低信噪比的三阶多项式相位信号参数估计问题提出一种新方法。该方法采用基于滤波器组的短时RadonWigner变换的时频检测算法完成信号的检测并获得信号瞬时频率的粗略估计，利用三阶多项式对瞬时频率进行最小二乘拟合，从而获得信号参数的粗略估计，再用局部二维搜索和快速傅里叶解线调的方法完成信号参数的精确估计。计算机模拟仿真证明了此方法在低信噪比和较短数据条件下的有效性。

Combination of instantaneous frequency curve fitting and local search method for 3-order polynomial phase signal parameter estimation

3 order polynomial phase signals (PPS) are important and common non-stationary signals. A new method was proposed to estimate the PPS parameters in low SNR under non-cooperative condition. Firstly, filter bank combining with short time Radon-Wigner transform algorithm was adopted to get a rough estimation of the instantaneous frequency(IF) curve. Then, a 3-oder polynomial was used to fit the IF curve by least square method and a rough estimation of the signal parameters was made. After that, local two-dimension searching and FFT dechirp method were used to get the parameter s accurate estimation. The merit of this new method is validated by illustrative examples in low SNR and relatively short data length condition.