Linear Prediction Theory: Mathematical Basis for Adaptive Systems

The theory presented in this work forms the basis of many algorithms for parameter estimation, adaptive system identification, and adaptive filtering. Linear prediction theory has applications in such fields as communications, control, radar and sonar systems, geophysics, estimation of economic processes, and training problems in synthetic neural nets. Emphasis is placed on three main areas. First, the mathematical tools required for the most important linear prediction algorithms are derived in a unified framework. Second, the relationships between different approaches are pointed out, thus allowing the selection of the optimal technique for a particular problem. Third, the material is presented in the context of results in algorithm research, with many references to publications in the field. The book is suitable for a graduate course on adaptive signal processing and should be useful for practising engineers faced with the problem of designing systems for operation in time-varying environments.