在脉冲多普勒体制雷达系统中,脉冲压缩后信号强度常常服从Gamma分布,使得诸如信噪比、信干噪比等指标的统计评估本质上是针对Gamma分布均值比进行的。 传统的Gamma分布均值比指标分析主要采用最大似然估计方法,不仅忽略了先验信息的使用,而且区间估计中的置信度也无法从概率意义上给出解释。为了克服这些缺点,文中通过推导Gamma分布均值比分布与Beta分布的关系,证明Gamma分布参数的后验分布为对数凹分布的特性,提出了基于Gibbs采样的Gamma分布均值比的贝叶斯统计推理框架,推导得到了后验点估计和区间估计方法,实测数据处理验证了算法的有效性。

In pulse Doppler radars, the pulse compressed signal intensity belongs to Gamma distributions, so that the evaluation of indices, such as signal-to-noise ratio and signal to interference plus noise ratio, is indeed the statistical inference of Gamma means ratio. The classical inference method for Gamma means ratio is maximum likelihood estimation; whereas, it not only neglects the use of prior information, but also results in unexplained confidence levels from probability views. To overcome these shortcomings,the posterior distribution of the Gamma parameters follows log-concave distribution is proved by deriving the relationship between the ratio′s distribution of Gamma means and the Beta distribution. Thus, the Bayesian statistical inference for Gamma means ratio based on Gibbs sampling methods is proposed, and posterior point estimation methods as well as interval estimation ones are proved. The effectiveness of the method has been verified by the measured data process.

TN973