You only need to keep track of the previous state. Key Takeaways for New Learners
( K_k = P_k / (P_k + R) ). Translation: How much do we trust the sensor vs. our model? If sensor noise ( R ) is small, gain ( K ) is high (trust sensor). If ( R ) is huge, gain ( K ) is low (ignore sensor). kalman filter for beginners with matlab examples by phil kim
Even with no motion, the Kalman filter is a superior averaging tool because it adapts – when it is uncertain, it trusts the measurement; when it is confident, it smooths the noise. You only need to keep track of the previous state