No label defined (Q5604977)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 3326154
Language Label Description Also known as
default for all languages
No label defined
    English
    No label defined
    scientific article; zbMATH DE number 3326154

      Statements

      0 references
      1969
      0 references
      The earlier chapters of this book deal with a general nonlinear continuous time estimation problem, and may be regarded as complementing the work of Kalman and Bucy on linear filtering. The author considers a (vector) message process which follows a first order nonlinear differential equation excited by a linear time varying function of white Gaussian noise. Observations are available of nonlinear functions of the message, corrupted by stationary white Gaussian noise (in-dependent of that generating the message). The problem is to obtain the minimum mean square error estimate of the message, based upon past observations. This is solved by developing a partial differential equation for the distribution of the message conditional upon past observations, using Bayes theorem together with an approach similar to that normally used to derive the Fokker-Planck equation. A differential equation is then deduced for the required estimate, which is the mean value of this conditional distribution. In order to obtain a practical solution, the author considers approximations obtained using expansions about the latest estimated message value, of the nonlinear functions in the message and observation model. The resulting differential equation can then be solved, given the covariance matrix of the message estimation error. This matrix is obtained as the solution of a coupled differential equation, which, to the order of approximation used, is similar to the matrix-Riccati equations of the linear filtering problem; it includes however, a further term involving second derivatives of the nonlinear function in the message model. The final two chapters of the book consider particular applications to analogue communication theory, with the aim of constructing realisable demodulators. Special solutions are given in the linear stationary case, using results for estimate error variances related to those of the Wiener-Hopf filtering theory. In other cases, the developed approximate solution to the nonlinear problem is applied, and the results of simulation studies given, which confirm the adequacy of the approximations. The material in this book is well presented. In particular, the results of the earlier chapters are summarised and interpreted before the derivations are given. The book fits in well with the aims of the MIT Research Monograph series.
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references