Integrated the CMSIS DSP library and finished the initial implementation of the Predict function

Josh Leverette's picture

I was able to successfully integrate the CMSIS DSP library into my project, in the form of the mbed-dsp package. My first attempt at this was not successful, due to some makefile issues, but taking an alternative approach was able to clear things up in (relatively) short order.

After getting that done, I now had the matrix functions I needed to properly manipulate my arrays. I have now written an initial implementation of the Kalman predict function, making heavy use of the CMSIS DSP library functions. The code is untested as of this moment, and I am concerned that upgrading my mbed library may cause unexpected, breaking changes in other parts of my code, just due to the nature of updating libraries attached to a project. I plan to make testing my current code (the new stuff and the new mbed version) part of my work tomorrow, but I expect to get other stuff done as well, hopefully. If so, there's every chance that by Monday I could have a functional (even if not properly tweaked) Kalman filter active in my code base! I find that prospect exciting.

GitHub link (kalman.h and kalman.cpp are where the current adventure is in progress)