Expand description
A crate providing an implementation of the unscented Kalman filter algorithm.
This implementation follows the steps presented in ‘this guide by mathworks’. It is able to support multiple different kinds of measurement updates through separation of the update and innovation steps of the filter.
Examples for the usage of this crate can be found in the examples folder.
TODO
- Change the filter methods to not take control. This can be handled in the provided closures instead. Is a more general interface for systems that do not specifically break out control.
- Support no_std
- Add some more examples: one with only one measurement update and one that does not require ode_solvers
Structs
Unscented Kalman filter structure. Contains the state of the filter and its covariance as well as a set of weights calculated for the algorithm based on the initialization parameters.
Type Definitions
Type for control vectors of dimension C
Type for covariance matricies of dimension SxS
Type for cross covariance matrices of dimension SxY
Type for measurement outputs of dimension Y
Type for state vectors of dimension S