一、序言 滤波的任务就是要从获得的信号与干扰中尽可能地滤除干扰,分离出所期望的信号。这个问题的解决,从维纳的开创性工作开始,得到了卡尔曼等人的推广和提高,目前仍然处于方兴未艾的阶段。但是任何将维纳滤波理论与卡尔曼滤波理论用于实际的人都知道,他们必须事先具备信号与干扰的某些统计知识,才能动手设计这类滤波器,而这些资料往往正好是缺少的。于是人们以极大的兴趣与精力转而研究如何根据观测数据来实时地调整滤波器的参数,使之适应实际情况,这就是自适应滤波问题。 通常所说的自适应滤波几乎都是在卡尔曼滤波模型上进行。采用不同的估值方法,根据观测数据,或者估计噪音的统 I. Preface The task of filtering is to filter out the interference as much as possible from the obtained signal and interference and separate the desired signal. The solution to this problem, beginning with Wiener’s pioneering work and getting the promotion and enhancement of Calman et al., Is still in the ascendant phase. But anyone who applies Wiener filter theory to Kalman filter theory knows that they must have some statistical knowledge of signals and disturbances in advance to design such a filter, which often happens to be missing. So people with great interest and energy to study how to adjust the filter parameters in real time according to the observed data to adapt to the actual situation, which is the adaptive filtering problem. Almost all of the so-called adaptive filtering is carried out on the Kalman filter model. Use different valuation methods, based on observed data or estimates of noise