论文部分内容阅读
Ｂｅｒｎｏｕｌｉ－Ｇａｕｓｓｉａｎ白噪声的检测存在于具有跳变输入的Ｋａｌｍａｎ滤波与最优平滑等问题中，邻近信号间的干扰和信号点数目无限增多是其中存在的两个主要问题．针对上述问题提出了一种基于最优平滑反卷积的Ｂｅｒｎｏｕｌｉ－Ｇａｕｓｓｉａｎ白噪声检测方法和新的多值信号建模方法，并基于固定区间平滑的频域性质和极大似然原理，给出了一个最优似然比检测指标．仿真表明，该方法具有好的分辨率，为存在波形重叠信号的高分辨检测提供了新的途径．
The detection of Bernouli-Gaussian white noise exists in the Kalman filtering and optimal smoothing with jump input. The two main problems are the interference between adjacent signals and the unlimited number of signal points. Aiming at the above problems, a new method of Bernouli-Gaussian white noise detection based on the optimal smoothing deconvolution and a new multi-valued signal modeling method are proposed. Based on the smoothing frequency domain property and the maximum likelihood principle of fixed interval, An optimal likelihood ratio detection index. Simulation shows that this method has good resolution and provides a new way for high resolution detection of overlapping waveform signals.