On the Modelling and Estimation of Communication Channels

Abstract

This chapter discusses the identification of communication channels characterized by a signal transformation and additive output noise. The chapter describes some general results regarding an abstract mathematical structure, called a ɛ-representation, in which channel identification problems can be set. With the use of this structure, the identification problems reduce in a certain way to estimation of parameters in a linear model. The chapter also presents a review of certain facts about linear estimators. The term communication channel is somewhat flexible in its common usage in radio engineering. Abstractly, however, a communication channel will always be a black box, a system that accepts inputs and produces outputs. It may provide a functional relationship between inputs and outputs, a stochastic relationship, or something that can conveniently be thought of as a little bit of both. A point-to-point communication system, as opposed to a communication network, is usually thought of as being made up of three separate successive components: (1) transmitter, (2) channel, and (3) receiver.