Numerical studies in nonlinear filtering (Q800676)

This is a book about state- and parameter estimation of nonlinear continuous-time systems working under a variety of conditions. The main purpose of the work is to provide finite-dimensional filters as approximate solutions to nonlinear state estimation problems, and to discuss their implementation. Theoretical analysis of the properties (convergence, accuracy, etc.) of the solutions presented is not the concern of this book. The following are the ten chapters of the book: 1. Preliminaries (containing a brief discussion of the Fujisaki-Kallianpur-Kunita filtering formula for calculating conditional expectations, which is central to all the later developments); 2. Estimation of parameters via state observation; 3. Filtering via Markov chain approximation; 4. A Kalman filter for a class of nonlinear stochastic systems; 5. Approximating filters for continuous-time systems with interrupted observations; 6. Estimation in a multitarget environment; 7. State and parameter estimation; 8. State estimation for systems driven by Wiener and Poisson processes; 9. Prediction via Markov chains approximation; and 10. Some extensions of linear filtering (covering the following problems: linear filtering with non-Gaussian initial conditions; estimation of maneuvering targets; detection problems; and state and covariance estimation). The book contains a large number of numerical studies. The behavior of every filtering, prediction, or parameter estimation algorithm introduced is illustrated by means of relatively simple but interesting numerical examples. This unique feature of this work makes it a useful reading for those who are engaged in the field of nonlinear filtering.