First glance at Kalman Filter
When I encountered Kalman Filter for the first time, I found it very complex and totally different from what I knew about filters before. Now I think I understand why I would feel it different.
1. Kalman filter is in time domain. Unlike low-pass filter, etc, Kalman filter is in time domain. This may be not so intuitive.
2. Kalman filter requires model (process model and measurement model). Even for a very simple problem, e.g., given a measurement sequence of a fixed length, we need to build up a model in order to apply Kalman filter. At first, I don’t understand. Now I see when we deal with very simple problem such as measuring a length, Kalman filter may be to complex to apply. But for many complex practical and theoretical problems, Kalman filter is very powerful.
3. Kalman filter works in a recursive manner. This is another problem that troubles me. At first, I don’t understand. Now I know this is a very good manner because it is time and memory saving.