import mvn
Example application: http://en.wikipedia.org/wiki/Talk:Kalman_filter#Example_Animation
- I was learning all of this as I did it, I may have made some mistakes.
- Often I was more interested in whether I could than whether I should.
- If it doesn't have automated tests, it probably doesn't work.
The goal is to make these probability distributions 'easy'. Not all of this works yet.
result = sensor1 & sensor2 & sensor3
# Linear Kalman Filter
state[t+1] = (state[t]*stm + noise) & measurment
# Unscented Kalman Filter
state[t+1] = mvn.Mvn(stateupdate(state[t].simplex())) & measurement
# Particle Filter
# (states are mixtures of points)
state[t+1] = (stateupdate(state[t]) & measurement).resample()
mix = Mixture([A,B,C])
mix = mix.fit(data)
M = mvn.Mvn(data)
M[0] = 10
T = np.Matrix(...)
dist = mvn.Mvn(data)
dist*T == mvn.Mvn(data*T)
dist2d = dist[:2]
p = dist.inBox(corner_1,corner_2)
mvn.Mvn(data).plot()