Contents
When to use the Kalman filter on time series?
A Kalman filter, suitable for application to a stationary or a non-stationary time series, is proposed. It works on time series with missing values. It can be used on seasonal time series where the associated state space model may not satisfy the traditional observability condition.
What is Kalman filter in time series?
Kalman filtering is an algorithm that produces estimates of unknown variables that tend to be more accurate than those based on a single measurement alone (sorry, I copypasted definition from wiki article). In other words, Kalman filter takes time series as input and performs some kind of smoothing and denoising.
Who is the creator of the Kalman filter?
FUN FACT: The Kalman filter was developed by Rudolf Kalman while he worked at the Research Institute for Advanced Study in Baltimore, MD. For the sake of introducing the Kalman filter, let’s take a simple model sometimes referred to as the “local level” model, which has a state equation of
Which is the state equation for the Kalman filter?
For the sake of introducing the Kalman filter, let’s take a simple model sometimes referred to as the “local level” model, which has a state equation of where we assume wt ∼ N (0,τ 2) w t ∼ N ( 0, τ 2) and vt ∼ N (0,σ2) v t ∼ N ( 0, σ 2).
How does the Simulink model work with Kalman filter?
The Simulink model contains two PI controllers for tracking the desired orientation and speed for the car in the ctrlKalmanNavigationExample/Speed And Orientation Tracking subsystem. This allows you to specify various operating conditions for the car and test the Kalman filter performance.
Why is the Q matrix chosen to be time varying?
The variance of the process noise w, the Q matrix, is chosen to be time-varying. It captures the intuition that typical values of are smaller when velocity is large. For instance, going from 0 to 10m/s is easier than going from 10 to 20m/s.