Filtering

A fingerprinting approach that increases accuracy. The filter is applied on top of the static (deterministic or probabilistic) approach.

Bayesian
a filtering approach using Bayes' rule, and usually what is meant by filtering in the context of location estimation. Noisy sensor data is used to update the belief that a system is in a certain state at a given point in time.

Kalman
The most common type of Bayes filter. A Kalman Filter (KF) is used in some filtering approaches to fingerprinting. PKF is a specific type of Kalman Filter, or possibly several, as it has been used to represent Position Kalman Filter, Particle Kalman Filter, Projective Kalman Filter, and at least a couple of others. Normally, a Kalman Filter is contrasted with a Particle Filter. Kalman Filters apply to normal distributions.

Particle
Particle Filters apply to multimodal distributions. Since Wi-Fi fingerprints are often multimodal, it makes sense that particle filters would be used over Kalman filters, though Kalman filters have been used for WiFi location estimation.