Informally, kalman filter helps us to improve estimation of data when measurement is subject to noise. 3 basic concepts to understand
Data collected, it is inaccurate due to noise and also lack of accuracy due to collection mechanism, ie. inaccurate sensor
typically denoted as Z
Our estimation of actual value, based on measure, generally it will be different from measurement, as we assert that measure is not accurate
Our prediction of next measurement, ie. T+1, it is based on the measurements and estimations we've made so far, ie. 0 - T