分量载频差极小时频信号的超分辨率分析方法

针对多个调制形式完全相同且分量间载频差极小的时频信号,提出了一种超分辨率分析信号的模型和方法。首先,以各分量的幅度和相位为状态变量,信号的时域观测和基于相同调制形式而构建的瞬时频率斜率为测量方程,为分量间载频差极小的时频信号构建多观测的状态空间模型。然后,为了估计模型需要的分量数目,利用匹配解调变换将原信号解调为近似平稳的窄带信号,再使用稀疏迭代自适应协方差估计该窄带信号在时间窗内的协方差,基于这一估计协方差并利用判别函数法估计该窗内的分量数目。最后,以多个滑动时间窗内估计结果的众数作为信号分量数目,就可基于提出的模型以及数据融合滤波算法对时频信号进行超分辨率分析。仿真结果表明,所提方法的性能优于其他方法。

Super-resolution analysis method for time-frequency signal with very small carrier frequency difference

A model and method for super-resolution analysis of time frequency signal with identical modulation forms and very small carrier frequency difference between its components is proposed. Firstly, the amplitude and the phase of each component are taken as the state variables, the time-domain observation of the signal and the instantaneous frequency slope based on the same modulation form are taken as the measurement equations, and the multi-observation state space model is constructed for the time-frequency signal with minimal carrier frequency difference between components. Then, in order to estimate the number of components required by the model, the original signal is demodulated into a nearly stationary narrow-band signal by using matched demodulation transform, and then the covariance of the narrow-band signal in the time window is estimated by sparse iterative adaptive covariance. Based on the estimated covariance, the number of component in the window is estimated by using the discrimination function estimator method. Finally, taking the mode of estimation results in multiple sliding time windows as the number of signal components, the super-resolution analysis of time-frequency signal can be carried out based on the proposed model and the data fusion filtering algorithm. The simulation results show that the performance of the proposed method is better than other method.