Pulling the information out of the clutter

Abstract

Fixed and wireless telecommunications systems, sonar systems, navigation devices, image processing algorithms-these are all examples of where signal processing is used. Much signal processing is based on statistical models of processes generating the signals and the contaminating noise. This paper traces the development of statistical processing theories, beginning with Wiener filtering, continuing through Kalman filtering, and ending with Hidden Markov Models. Different assumptions underpin these theories, and also very different mathematics. Yet a number of common features remain.