Tracker Component Library Release 4.0, January 2019 https://github.com/USNavalResearchLaboratory/TrackerComponentLibrary

A paper describing a number of features of the library is
D. F. Crouse, "The Tracker Component Library: Free Routines for Rapid Prototyping," IEEE Aerospace and Electronic Systems Magazine, vol. 32, no. 5, pp. 18-27, May. 2017.

These are the release notes for the version 4.0 of the Tracker Component Library. The Tracker Component Library is a collection of Matlab routines for simulating and tracking targets in various scenarios. Due to the complexity of the target tracking problem, a great many routines can find use in other areas including combinatorics, astronomy, and statistics.

Making this code available and open source is in the spirit of OMB M-16-21, which can be viewed online at: https://sourcecode.cio.gov/

Those looking for magnetic field synthesis code might want to look at ./Sample Code/Magnetic Models/ .

As of version 3.0, the library has been split into two parts. This is the publicly available part. A number of functions have been placed into a supplement that is only available to the U.S. Government Agencies and their contractors.The limited distribution supplement can be downloaded by authorized individuals from di2e.net at: https://bitbucket.di2e.net/projects/TCL

Those looking to get a quick idea of a very simple end-to-end tracking algorithm (track initiation, data association, maintenance, and termination) might want to look at ./Sample Code/Basic Tracking Examples/demo2DIntegratedDataAssociation.m to see an example of a complete GNN-JIPDAF tracker run on a simple 2D tracking scenario, which can be easily transformed into a normal JIPDAF by changing a single parameter. The tracker is made from the modular components of the library. A lot of other sample code is also provided to demonstrate the use of other parts of the library.

To use the full library, add the library and all of its subfolders to your active path in Matlab. Some functions are available as C/C++ files for use in Matlab either because they use third-party libraries (and must be compiled to be used) or because the native Matlab implementations provided are too slow in certain circumstances. These files can be compiled by running the CompileCLibraries function. Precompiled code is not distributed with the library. Note that a C/C++ compiler supported by Matlab must be installed. See below for comments regarding compilation.

NOTABLE CHANGES SINCE VERSION 3.5

• Routines related to stochastic calculus have been changed and new ones have been added. They are in: Mathematical Functions/Statistics/Stochastic Processes/
• New continuous-time state propagation models based on Itô-Taylor expansions have been added to: Dynamic Estimation/State Propagation/Continuous Time/
• All continuous-time dynamic models in the library have been updated to work with these (and bugs fixed) and some new dynamic models have been added.
• Functions splitting the measurement prediction and update steps for most filters have been added. This makes hypothesis generation faster, particularly with cubature filters.
• Functions related to robust estimation have been added to: Mathematical Functions/Statistics/Robust Statistics/ One can, for example, fit lines to data corrupted with arbitrary bad measurements.
• Added many more combinatoric functions and updated existing ones. See Mathematical Functions/Combinatorics/
• Joint matrix diagonalization algorithms (useful for independent component analysis) have been added. See Mathematical Functions/Basic Matrix Operations/Joint Matrix Diagonalization/
• Misc. statistical functions. For example, unbiased estimates of higher- order moments and cumulants are difficult to derive, but can be found using the unbiasedMomentCumulant function.

NOTABLE CHANGES SINCE VERSION 3.01

• More compiled versions of 2D and 3D assignment code are available and the C implementations have been more modularized to make then simpler to remove from the Matlab parts of the library and use elsewhere.
• Sped up the dense interior point linear programming method by a large factor compared to the textbook implementation.
• A few new optimization methods.
• Improved the probability distributions. Added new ones, such as the wrapped normal distribution, and added entropy to many of them.
• For the many functions that have no Matlab implementation and must be compiled to be used (mostly those using third-party libraries), shell .m files have been added to the library, so that one can type "help" and the name of the function to get information on how to use it. Also, if the function is called without having compiled it, then an error alerts the user to the need to compiled the function.
• Bug fixes and small changes.
• Updated third-party code. Removed reliance on third-party functions for line searches.

NOTABLE CHANGES SINCE VERSION 3.0

• Minor corrections.
• Minor modifications to files to support the final release of the limited distribution supplement to the library.
• New 2D and 3D assignment algorithms, among other additions.

NOTABLE CHANGES SINCE VERSION 2.5

• Many new functions have been added.
• Many corrections have been made and a number of functions in different areas have been added or made more versatile/ stable.
• A number of routines related to signal processing have been added to the library in the Mathematical Functions/Signal Processing folder.
• Less reliance on third-party libraries and compiled code. NASA's SPICE library has been removed and the function readJPLEphem can be used to read in NASA JPL's ephemerides without compiling anything. Additionally, the directGeodeticProb function has been rewritten in Matlab and does not use GeographicLib, though the GeographicLib library is still used to solve the indirect geodetic problem.
• Spherical harmonic synthesis has been simplified and now works with vector coefficients and complex coefficients. Additionally, in addition to obtaining a potential and a gradient, one can obtain a Hessian matrix, opening the door to simulation of gravity gradiometer measurements. Fitting of low to moderate order spherical harmonic coefficients to measured data is now possible with the fitSpherHarmonics function.
• Arbitrary-order multivariate B-spline fitting, evaluation, differentiation, and integration has been added.
• Linear dynamic models expressed in non-Cartesian coordinate systems have been added along with conversions of states between coordinate systems. See /Dynamic Models/Continuous Time/Non-Cartesian Position/ and /Coordinate Systems/State Coordinate System Conversion/

COMPILED CODE:

The compilation of the library has been tested under Matlab2018b for Mac OS X 10.12 using the Matlab-supported compiler offered free by Apple. Under Linux, gcc was used as the compiler. The code should also compile under Windows 10 using minGW. Precompiled code is not distributed with the library.

EXTERNAL SOLVERS:

Almost all of the function in the library work without any external resources. However, the folder ./Mathematical Functions/Polynomials/Generic Multivariate Polynomials/ contains the function solvePolySysWithExtProg. This is an interface to external solvers for simultaneous multivariate polynomial systems. The solvers are external programs and are not included. Though the functions polyRootsMultiDim and solveQuadBivarEq can solver certain types of simultaneous multivariate polynomials, they are limited. For more general systems, an external solver is the best choice. The functions polyMeasConvert, polyMeasConvertAsync, and DopplerOnlyInit6D are located in subfolders entitled 'Uses External Solver" and use the function solvePolySysWithExtProg. Supported external solvers for the function solvePolySysWithExtProg are PHCpack, Bertini, and the certified solver in Macaulay2. Note that Macaulay2 tends to fail more often than the other solvers.

The solvers can be downloaded from: PHCpack for Mac OS X, Windows and Linux: http://homepages.math.uic.edu/~jan/download.html Bertini (version 1.5) for Mac OS X and Linux and Windows, only with Cygwin: https://bertini.nd.edu/download.html Macaulay2 for Mac OS X, Linux, and Windows: http://www.math.uiuc.edu/Macaulay2/

For the external solvers to work, they need to be added to the search path used in Matlab so that the phs/bertini/m2 commands can be called using the system command in Matlab.

DATA SOURCES:

If the library was downloaded without associated data files for solar system ephemerides as well as magnetic, gravitational and terrain models, then the relevant data files will have to be downloaded The original sources of the data are:

1. Earth2014 terrain model: https://geodesy.curtin.edu.au/research/harmonic/ The files Earth2014.BED2014.degree2160.bshc, Earth2014.ICE2014.degree2160.bshc, Earth2014.RET2014.degree2160.bshc, Earth2014.SUR2014.degree2160.bshc, Earth2014.TBI2014.degree2160.bshc should be placed into ./Terrain/data . Though data files to degree and order 10,800 are also available, they are not currently supported as the function spherHarmonicEval does not use extended precision arithmetic, which is required to avoid overflows when using the model.

2. EGM2008 terrain model (The DTM2006.0 model): http://earth-info.nga.mil/GandG/wgs84/gravitymod/egm2008/first_release.html The file Coeff_Height_and_Depth_to2190_DTM2006.0 should be placed into ./Terrain/data . The first time using the getEGM2008TerrainCoeffs function will be significantly slower than subsequent calls, assuming the full model is loaded the first time.

3. The EMM2017 magnetic field coefficients: https://www.ngdc.noaa.gov/geomag/EMM/ The coefficients are included with the software that is provided. The files having names like EMM2015.COF and EMM2000.COF should all be zipped into an archive with the name EMM_Coefficients.zip and placed in ./Magnetism/data . The files with names like EMM2015SV.COF andEMM2000SV.COF should all be zipped into an archive with the name EMM_Secular_Variations.zip and placed into ./Magnetism/data . If all of the data for a single year is loaded, then a .mat file containing the data is placed in the data folder to speed up subsequent calls to the function.

4. The IGRF12 magnetic field coefficients: http://www.ngdc.noaa.gov/IAGA/vmod/igrf.html The file http://www.ngdc.noaa.gov/IAGA/vmod/igrf12coeffs.txt can be downloaded and placed into the folder ./Magnetism/data .

5. The WMM2015v2 magnetic field coefficients: https://www.ngdc.noaa.gov/geomag/WMM/soft.shtml The coefficient file WMM2015v2COF.zip shown be downloaded, unzipped and the file WMM.COF placed into ./Magnetism/data .

6. EGM2008 gravitational coefficients http://earth-info.nga.mil/GandG/wgs84/gravitymod/egm2008/egm08_wgs84.html Download the file EGM2008_to2190_TideFree.gz, which is under "EGM2008 Tide Free Spherical Harmonic Coefficients". Decompress the file and place it in ./Gravity/data .

7. EGM96 gravitational coefficients http://earth-info.nga.mil/GandG/wgs84/gravitymod/egm96/egm96.html The file egm96.z should be downloaded and decompressed, resulting in a file names egm96 without an extension, which should be placed in ./Gravity/data .

8. EGM2008 parameters needed to convert height anomalies to geoid undulations: http://earth-info.nga.mil/GandG/wgs84/gravitymod/egm2008/egm08_wgs84.html The "Correction Model" file Zeta-to-N_to2160_egm2008.gz should be downloaded, ungzipped and placed in ./Gravity/data .

9. EGM96 parameters needed to convert height anomalies to geoid undulations: http://earth-info.nga.mil/GandG/wgs84/gravitymod/egm96/egm96.html The file corrcoef.z should be decompressed and placed in ./Gravity/data .

10. NASA JPL's DE430t planetary/solar/lunar ephemerides ftp://ssd.jpl.nasa.gov/pub/eph/planets/Linux/de430t The file linux_p1550p2650.430t should be downloaded and placed in ./Astronomical Code/data for use by the readJPLEphem function. The readJPLEphem function can read other NASA ephemerides, but the DE430t set are the default.

11. NASA JPL's GL0900C gravitational model of the Moon http://pds-geosciences.wustl.edu/missions/grail/default.htm Download the file jggrx_0900c_sha.tab and place it in ./Gravity/data .

12. The FES2004 tide model of the effects of ocean Earth tides. http://maia.usno.navy.mil/conv2010/convupdt/convupdt_c6.html The file fes2004_Cnm-Snm.dat should be downloaded and placed in ./Gravity/Tides/data .

13. The Hipparcos 2 Star Catalog http://cdsarc.u-strasbg.fr/viz-bin/Cat?I/311 Place the file hip2.dat in ./Astronomical Code/data .

14. A low-resolution image of the "Blue Marble" Earth from NASA with land, ocean color, sea ice and clouds. https://visibleearth.nasa.gov/view.php?id=57735 Place the file in Misc/data . This is used as the default map to show on the Earth given by the plotMapOnEllipsoid function.

January 2019 David F. Crouse, Naval Research Laboratory, Washington D.C.
(UNCLASSIFIED) DISTRIBUTION STATEMENT A. Approved for public release.

LICENSE:

The source code is in the public domain and not licensed or under copyright. The information and software may be used freely by the public. As required by 17 U.S.C. 403, third parties producing copyrighted works consisting predominantly of the material produced by U.S. government agencies must provide notice with such work(s) identifying the U.S. Government material incorporated and stating that such material is not subject to copyright protection.

Derived works shall not identify themselves in a manner that implies an endorsement by or an affiliation with the Naval Research Laboratory.

RECIPIENT BEARS ALL RISK RELATING TO QUALITY AND PERFORMANCE OF THE SOFTWARE AND ANY RELATED MATERIALS, AND AGREES TO INDEMNIFY THE NAVAL RESEARCH LABORATORY FOR ALL THIRD-PARTY CLAIMS RESULTING FROM THE ACTIONS OF RECIPIENT IN THE USE OF THE SOFTWARE.