ABSTRACT
The connotation of "maximum likelihood" is a setting in which nothing is known α priori about the unknown quantity, but there is prior information on the measurement process itself. There are many ways to estimate an unknown quantity from available data. This chapter presents mean-square estimation and the Wiener filter. A discussion of recursive estimation is included, since it provides a good background for the Kaiman filter. The techniques presented can be described as classical techniques in that they represent the state of the art prior to 1960, when Kaiman and Bucy developed their approach. Mean-square estimation is a special case of Bayesian estimation. The accurate determination of the noise parameters is a crucial part of the estimation process.
