kalman_filter

Goal

this is the implementation of Kalman filter

Learning Resources

https://www.youtube.com/watch?v=CaCcOwJPytQ&index=1&list=PLX2gX-ftPVXU3oUFNATxGXY90AULiqnWT https://www.avrfreaks.net/forum/tut-theory-what-kalman-filter?name=PNphpBB2&file=viewtopic&t=84846

What to run and How to run

there are three python script files, each file can be run seperatly, the file with main logic calculation is in kalman_filter_impl.py

1. temperature.py corresponds to the first few lectures, that walks through the single dimension senario
2. kalman_filter_youtube.py contains all the main logics (8 steps process) from the youtube videos. when script is run, you are expected to see a drawing with three plots. one for predicted states(blue), one for measurements (read), one for adjusted states(green). As shown in the lectures, three interations are done in this implementation as well, and the result are verified. (the calculation for the last itration from the video has some errors, and you should be able to find out)

3. matrix_test.py contains some matrix operator tests, plus assertion of the what happens in each step in the complete pipeline.
4. kalman_filer_upgrade.py adds things on top of kalman_filter_youtube.py. two extra parameters, noise, time series data generation.

differece between this and the requirement from homework

1. the prediction equation is different, because it takes acceleration into account.
2. no noice is taken into account.
3. the state matrix only contains 2 variables, while requirement has 6. The the only difference should be so that the A_Matrix, H and H_Tranpose matrix are different.