摘要
利用小波包分析将信号变换到其小波包系数域上 ,通过研究平稳随机噪声在各子空间上小波包系数的统计特性 ,结合 Robust检测理论的相关结论 ,对非高斯噪声下的信号检测问题建立了统一的框架 ,并提出了一种新的检测算法。仿真实验表明 ,此算法不仅具有良好的鲁棒性和广泛的适用性 ,而且能够更为充分地利用噪声统计分布信息 ,从而有效地提高检测效率。
In some practical problems, information about the distribution of background noise is not known precisely, when the application of classic signal detection theory is greatly limited. Using wavelet packet, this paper transits signal to the domain of wavelet packet coefficients. By studying the statistic property of wavelet packet coefficients of stationary noise at each subspace, combing the relative results of robust detection theory, a unified frame work is built to signal detection in additive arbitrary distribution non-Gaussian noise. Furthermore, a new detection algorithm is presented. Simulational experiment shows that the algorithm has good robustness and extensive applicability can efficiently exploit the statistic information of noise, so higher performance is achieved.
出处
《数据采集与处理》
CSCD
2003年第3期302-305,共4页
Journal of Data Acquisition and Processing
