Scribd
Upload
39 views2 pages

A Macsyma/Expert System For Nonlinear Filtering: G.L. Blankenshlp, T A. Lavlgna, T D.C. Macenany, T Yant

This document describes a prototype expert system written in MACSYMA for solving certain nonlinear filtering problems symbolically. The system accepts user input describing signal and measurement processes, analyzes the problem symbolically to compute equations like the Zakai equation, and generates FORTRAN code for numerical solutions. It can also test the well-posedness of filtering problems and generate code for computing likelihood ratios. Sample sessions are presented to illustrate how it works. Plans for further developing the system are also discussed.

Uploaded by

Subhashree Mohanty
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
39 views2 pages

A Macsyma/Expert System For Nonlinear Filtering: G.L. Blankenshlp, T A. Lavlgna, T D.C. Macenany, T Yant

This document describes a prototype expert system written in MACSYMA for solving certain nonlinear filtering problems symbolically. The system accepts user input describing signal and measurement processes, analyzes the problem symbolically to compute equations like the Zakai equation, and generates FORTRAN code for numerical solutions. It can also test the well-posedness of filtering problems and generate code for computing likelihood ratios. Sample sessions are presented to illustrate how it works. Plans for further developing the system are also discussed.

Uploaded by

Subhashree Mohanty
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

A MACSYMA/EXPERT SYSTEM FOR NONLINEAR FILTERING

G.L. Blankenshlp,t A. LaVlgna,t D.C. MacEnany,t


J.P. Quadrat,tt 1. Yan t

Abstract: A prototype "expert" system for the treatment of certain nonlinear filtering problems
is described with illustrative examples. The system is written in MACSYMA. It accepts user
input in symbolic form; it carries out the basic analysis of the user's problem in symbolic form
(e.g., computing the Zakai equation and the estimation Lie algebra for nonlinear filtering prob-
lems); and it produces output in the form of automatically generated FORTRAN code for the
final numerical reduction of the problem, e.g., FORTRAN code for the conditional statistics of
the Signal given the observation process. The system is also capable of testing the well-posedness
of a limited class of filtering problems by setting up the robust form of the Zakai equation and
applying some known theorems for well-posedness of that system. This is done in symbolic form
directly in the MACSYMA code. In addition, the system has a module Which generates FOR-
TRAN code for the computation of the likelihood ratio arising in connection with the filtering
problem. This code implements a sophisticated approximation for the Feynmann-Kac (or
Kallianpur-Streibel) representation of the conditional density in the estimation problem. Sample
terminal sessions are presented to illustrate its operation. The status of the system and plans for
its further development are described.

'ElectrIcal Engineering Department, University of Maryland, College Park, Maryland 20742. The research of G.L.
Blankenship and I. Yan was supported In part by NSF Grant ECS-83-15Q65. A. LaVlgna and D.C. MacEnany were
supported in part by ONR Graduate Fellowships. Address correspondence concerning this paper to G.L. Blankenship.
ttlNRIA, Domain. de Voluceau, Rocquencourt, B.P. 105,78153 LE CHESNAY CEDEX, FRANCE.

17
S. Albeverio et al. (eds.), Stochastic Processes in Physics and Engineering, 17-61.
© 1988 by D. Reidel. Publishing Company.
18 G. L. BLANKENSHIP ET AL.

1. Introduction

Nonlinear filtering theory has undergone a period of intense study over the past five years.

Powerful and sophisticated mathematical tools, including constructions from differential geometry,

stochastic partial differential equations, functional integration, stochastic mechanics, quantum

physics, and asymptotic analysis, among others have been brought to bear on the basic problem

of constructing recursive estimates of a signal process from measurements of a related process.

The anthology [IJ covers many of the key ideas, and the papers [2, 3J are excellent expositions of

much of the recent research. Because most of this work has employed sophisticated tools not in

the lexicon of most practicing engineers, and because much of the work has been expressed in the

theorem - proof format which has become the standard in control theory, it is likely that the

impact of this work on engineering design will be marginal at best, at least in the short run.

In this paper we describe a software system based on MACYSMA1 which embodies several

elements of the theory of nonlinear filtering, espeCially as described in [3J and [4, 5, 6, 7, 8J. The

system is easy to use, requiring only user specification of signal and measurement process equa-

tions in symbolic form and responses to a few simple questions on whether the model involves

vector or scalar signals and whether or not there are small parameters in the model. The system

computes the stochastic partial differential equation which describes the evolution of the condi-

tional density of the signal given the history of the observations; if smallparamters are present in

the model, it does an asymptotic analysis of this equation; it computes the differential geometric

structure of the approximating system; producing intermediate expressions which describes this

structure; and, as its final output, it writes a FORTRAN code for numerical computation of the

conditional mean of the signal. The FORTRAN code is generated "automatically" for each (sym-

bOUc) specification of the system model. The system is capable of checking the well-posedness of

the conditional density eqtlation using the criteria which have been derived for this purpose, e.g.,

in {6J. In addition, the system has a module which generates FORTRAN code for the computa-

tion of the likelihood ratio ariSing in connection with the filtering problem. This code implements

lMACSYMA Is a language for symbolic manipulation developed at Project MAC at MIT. MACSYMA Is a trade-
mark of Symbolics, Inc.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy