Implemented filters so far:
- Kalman filter
- Extended Kalman filter
- Second-order extended Kalman filter
- Unscented Kalman filter
Please use cmake to build all the codes.
The steps to compile are:
mkdir build
cd build
cmake ..
make
In a Windows system, a Visual Studio solution file (VS 2019) is provided.
This library utilizes Armadillo. In Windows system, the armadillo library is provided in "windows-libs" folder. The contents of windows-libs.zip need to be first extracted. Armadillo itself is very easy to use. More information on the Armadillo can be found here.
By default, now kalman-cpp uses blas and lapack. For Windows machine, working with blas and lapack is a messy stuff. Thus, we will use the precompiled blas and lapack from: https://www.fi.muni.cz/~xsvobod2/misc/lapack/.
The precompiled blas and lapack libraries are included in windows-libs.zip. There are four LIB files. Additionally, in "bin" folder, there are four corresponding DLL files as well. There are four files because two files are for the 32-bit platform, and the other two files are for the 64-bit platform.
The compiled binary must always be located in the same folder as these DLL files.
MATLAB m-files for each example are provided in 'm-files' folder. Octave can also be used instead of MATLAB.