一、引言在许多信号检测场合,信号是用一组参数(例如振幅、相位、波形)来描述的;而观察者不一定准确地知道这些参数。在本文中,我们讨论信号检测问题,其中允许信号和噪声这二者之一或二者的参数以不定的方式随时间而变化。这种时间变量的含义是:假定各参数可以用马尔可夫随机过程来模拟。处理这类问题的方法有若干种,有一般的和特殊的。让我们简单地讨论两种方法。解决这类问题的一种常用方法是:对每个未知参数作某种估计,并将这些估计值当作是真正的参 I. INTRODUCTION In many cases of signal detection, the signal is described by a set of parameters (eg amplitude, phase, waveform); the observer does not necessarily know these parameters exactly. In this paper, we discuss the problem of signal detection where parameters of either or both of signal and noise are allowed to vary over time in an indefinite manner. The meaning of this time variable is: Suppose each parameter can be simulated by Markov random process. There are several ways to deal with this type of problem, both general and special. Let’s briefly discuss two methods. A common method of solving this type of problem is to make some estimate of each unknown parameter and treat these estimates as true