Swift implementation of Conventional Kalman Filter algorithm
Kalman filter is a widely applied algorithm to get a more accurate guess in noisy environment. It has a lot of applications in real life such as guidance, navigation control for vehicles, etc. Although it is mostly used to filter GPS data, this framework doesn't have a ready-to-use solutions that work with GPS and is more general implementation of algorithm.
Kalman filter can work with anything that adopts KalmanInput protocol. Framework provides Matrix
struct that conforms to this protocol, although you can use anything that is more suitable for you. For
example, framework also provides Double's extension with KalmanInput and you can use it if your
KalmanFilter has only 1 dimension.
The code below is the example of usage of 1D KalmanFilter taken from
here.
let measurements = [0.39, 0.50, 0.48, 0.29, 0.25, 0.32, 0.34, 0.48, 0.41, 0.45, 0.46, 0.59, 0.42]
var filter = KalmanFilter(stateEstimatePrior: 0.0, errorCovariancePrior: 1)
for measurement in measurements {
let prediction = filter.predict(1, controlInputModel: 0, controlVector: 0, covarianceOfProcessNoise: 0)
let update = prediction.update(measurement, observationModel: 1, covarienceOfObservationNoise: 0.1)
filter = update
}All the names of properties and methods' parameters names are taken from wikipedia page.
If you are looking for a good source to understand how Kalman Filter works then take a look at
this page,
this one and there is a great series of
video tutorials on udacity