Identification of Time-Varying Processes in Signal Processing
Maciej Niedzwiecki, Technical University of Gdansk, Poland
0471 98629 1
June 2000
Hardback
340pp
Time Varying Process Identification (TVPI) is the latest advance in the evolution of Adaptive Signal Processing. TVPI has many applications in adaptive control, and its techniques facilitate adaptive noise reduction, echo cancellation and predictive coding of signals in communication systems.
Authored by an internationally recognized authority in the field, this unique and highly topical volume addresses the identification of time-varying characteristics of dynamic processes in signal processing. It features numerous case studies, including restoration of archive gramophone recordings and design of adaptive auto-pilots for ships.
Includes detailed theoretical material illustrated with experimental results
Presents an invaluable assessment of the estimation memory and bandwidth of different identification algorithms
A dynamic research tool for researchers and postgraduate students of adaptive signal processing
Contents:
Introduction:
Mathematical Models of Non-stationary Signals and Systems
The Concept of Estimation Memory
The Local Estimation Approach to Identification of Time-Varying Systems:
Estimation based on Process Segmentation
Weighted Least Squares Estimators
The Least Mean Sqaures Family
Local Estimation Algorithms for ARMA Processes
The Basis Functions (BF) Approach to Identification of Time-Varying Systems:
Estimation based on Process Segmentation
Weighted Basis Function Estimators
The Kalman Filtering (KF) Approach:
Algorithms Based on Kalman Smoothing
Adaptive Memory Tuning and Computational Safeguards:
Adaptive Memory Tuning Procedures
Identifiability Issues and Regularized Techniques
Special Techniques and Case Studies:
Identification of Time-Varying Processes in the Presence of Measurement Noise and Outliers
Identification in the Presence of Feedback.
Series: Wiley Series in Telecommunications and Signal Processing
