基于高阶统计量的信号特征描述

Higher-Order Statistics for Signal Characterization

在特征评估或者模型验证中,信号统计特征是完整描述和预测非线性系统输出的关键。传统的描述信号特征的方法是基于它的功率谱密度PSD和概率密度函数PDF,有其局限性。通过介绍无线电信号的高阶统计量计算,估算和修正信号的高阶自相关,表明对现代的信号处理研究,基于高阶统计量计算实际中的微波信号是可行的,可以完整描述信号特征。

In a performance evaluation test or model identification experiment,signal statistical characteristics is the key to properly interpreting and predicting the outcome of a nonlinear system.The traditional signal characterization is based on its power spectral density and probability density function,but the method has its limitation.This paper presents the implementation for the computation of higher-order statistics of a radiosignal,the measurement（and thus the correction） of the signal＇s higher order auto-correlations,thus states the feasibility for the higher order statistics computation of practical microwave signals,and that the high-order statics could completely describes the signal characteristics.