The book is structured into five logical parts that build in complexity: dandelon.com Part I: Recursive Filter:

By next week, you will be the person on the forum explaining the Kalman Gain to other beginners.

If your GPS is cheap and noisy, the filter trusts the prediction more.

In every case, the core idea is the same as Phil Kim’s MATLAB examples: predict, measure, correct, repeat.

where A is the state transition matrix, and w is the process noise.