# Module kalman

source code


Modules and misc. code related to the Kalman Filter.

Kalman filter algorithm as presented in "Probabilistic Robotics"

x_t is the state

u_t is a control vector

z_t is the observation vector

\epsilon_t is a random noise term with zero mean and covariance R_t.

\delta_t is a random noise term with zero mean and covariance Q_t.

state (x_t) evolves according to

x_t = A_t x_{t-1} + B_t u_t + \epsilon_t

Observation z_t is made according to

z_t = C_t x_t + \delta_t

Assume that the distribution over initial states is a Gaussian.

With these linear/Gaussian assumptions, the belief about the state all times t is Gaussian, so
we can represent it compactly by the mean (mu) and the covariance (sigma).


 Classes
KalmanModule