论文部分内容阅读
在作最佳检测设计时,杂波统计特性必须作为复随机过程来研究。本文考虑了把复杂波模拟为球面不变随机过程(SIRP),即假定它的PDF能表达为非负定二次型、高斯过程的一般形式。分析了SIRP的有关特性,表明它应遵循的一些基本要求,如正交分量联合PDF的圆对称性,或相当于相位分布的均匀性。对包络分布必须施加许用性限制。但大多数通常所采用的包络分布,包括威布尔,污染瑞利和K分布都是许用的。虽然一般的SIRP不是各态历经的,但最后提出了SIRP在集内扫描的杂波过程,恢复了各态历经性,也讨论了依据已经提出的合成扫描模型对此模型所进行的解释。
When making the best detection design, clutter statistics must be studied as complex random processes. In this paper, the complex wave is modeled as a Spherical Invariant Stochastic Process (SIRP), which assumes that its PDF can be expressed as a non-negative definite quadratic Gaussian process. The relevant characteristics of SIRP are analyzed, indicating some basic requirements that it should follow, such as the circular symmetry of the quadrature component joint PDF, or equivalent to the uniformity of the phase distribution. Entropy limits must be imposed on the envelope distribution. However, most of the commonly used envelope distributions, including Weibull, Contaminated Rayleigh and K distribution are permissible. Although the general SIRP is not ergodic, a clutter sweep of SIRP in the set is finally proposed and the ergodicity of the SIRP is restored. The model interpretation is also discussed based on the proposed synthetic swept model.