My simple attempt at writing a EKF/UKF in Haskell.
- stack build
- stack exec CannonBall
- open CannonBall.svg
Simulates a cannon ball shot with Newtonian equations of motion. The nonlinearity here is added to the measurement values. Note I think I have an issue with the UKF as the estimate seems to diverge after a long period. Looks like there might be some numerical instability with cholesky decomposition.
Simulates the Van Der Pol oscillator. WGN is added to both the sensor and the process