PreviousNext

 \newcommand{\R}[1]{{\rm #1}}  \newcommand{\B}[1]{{\bf #1}}
ckbs-0.20130204.0: Constrained/Robust Kalman-Bucy Smoothers
One web page per Section   All as one web page
Math in Latex   ckbs.htm _printable.htm
Math in MathML   ckbs.xml _priintable.xml

Purpose
The programs in this package compute a smoothed estimate of a time series given process models for the state evolution, measurement models for the observations, variance estimates, and the sequence of observations.

Affine Constrained Smoother
The program constrained_affine performs Kalman smoothing when the process and measurement models are affine. Affine inequality constraints may also be imposed on the state.

Nonlinear Constrained Smoother
The program constrained_nonlinear performs Kalman smoothing for general nonlinear process measurement. Nonlinear inequality constraints may also be imposed on the state.

Affine L1 Robust Smoother
The program robust_affine performs robust Kalman smoothing when the process and measurement models are affine and the measurement sequence may contain outliers.

Nonlinear L1 Robust Smoother
The program robust_nonlinear performs robust Kalman smoothing for general nonlinear process and measurement models when the measurement sequence may contain outliers.

General Student's T smoother
The program t_general performs robust Student's t Kalman smoothing for general nonlinear process and measurement models. The measurements may contain very large outliers. Also, there may be sudden changes in trend data. The user can specify which components of process and measurement residuals to model using Student's t.

MathML
This documentation is in XHTML + MathML. If you have trouble viewing it in your browser see Viewing MathML on the web page
     
http://www.seanet.com/~bradbell/packages.htm

System Requirements
This constrained Kalman-Bucy smoother runs under both the Matlab® and Octave While it will work under MS windows, these download instructions are for the Unix operating system.

Download

Release Tarballs
You can download a tarball release of ckbs from the directory
     
http://www.coin-or.org/download/source/CoinBazaar
Theses releases are named
     ckbs-0.
yyyymmdd.r.ext
where yyyy is the year, mm is the month, dd is the day, r is the release number, and ext is tgz or zip.

If ext is tgz the command
     tar -xvzf ckbs-0.
yyyymmdd.r.tgz
will create the ckbs-0.yyyymmdd.r directory.

If ext is zip the command
     unzip ckbs-0.
yyyymmdd.r.zip
will create the ckbs-0.yyyymmdd.r directory.

Using Subversion
If you are going to use subversion to download a version of ckbs, start with the following commands:
     mkdir ckbs
     cd ckbs
     web="https://projects.coin-or.org/svn/CoinBazaar/projects/ckbs"
For each release listed in the download directory you can use subversion to download the release with the commands
     svn checkout $web/releases/0.
yyyymmdd.r
This will create the ckbs/0.yyyymmdd.r directory. You can also download the most recent version of ckbs (called the trunk) using the commands
     svn checkout $web/trunk
This will create the ckbs/trunk directory.

Test Install
Change into the directory ckbs/0.yyyymmdd.r/example or ckbs/trunk/example and use Matlab® or Octave to run the program all_ok.m. It will print the test results for all the components of ckbs. You can run some more tests by changing into the directory ckbs/0.yyyymmdd.r/test or ckbs/trunk/test and running the program all_ok.m which will have a similar output.

Documentation
If you download a release, a copy of the documentation for the release can be found be viewed in your browser by opening the file
     ckbs/0.
yyyymmdd.r/doc/ckbs.xml
If you download the latest version (the trunk), a copy of the documentation for this version can be built by executing the commands
 
	./build_doc.sh
in the ckbs/trunk directory. You can then view the documentation in your browser by opening the file
     ckbs/trunk/doc/ckbs.xml

Contents
_contentsTable of Contents
licenseYour License to use the ckbs Software
ckbs_t_generalThe General Student's t Smoother
ckbs_nonlinearThe Nonlinear Constrained Kalman-Bucy Smoother
ckbs_L1_nonlinearThe Nonlinear Constrained Kalman-Bucy Smoother
ckbs_affineConstrained Affine Kalman Bucy Smoother
ckbs_affine_singularSingular Affine Kalman Bucy Smoother
ckbs_L1_affineRobust L1 Affine Kalman Bucy Smoother
utilityckbs Utility Functions
all_ok.mRun All Correctness Tests
whatsnewChanges and Additions to ckbs
wishlistList of Future Improvements to ckbs
bibBibliography
_referenceAlphabetic Listing of Cross Reference Tags
_indexKeyword Index
_searchSearch This Web Site
_externalExternal Internet References

Input File: ckbs.omh