Influence of a non-Gaussian state model on the position estimation in the nonlinear filtration
Stanislaw Konatowski, Barbara Pudlak, Barbara Pudlak · Proceedings of SPIE, the International Society for Optical Engineering/Proceedings of SPIE · 2007
In navigation systems with nonlinear filtration algorithms extended Kalman filter is being used to estimate position. In this filter, the state model distribution and all relevant noise destinies are approximated by Gaussian random variable. What is more, this approach can lead to poor precision of estimation. Unscented Kalman filter UKF approximates probability distribution instead of approximating nonlinear process. The state distribution is represented by a Gaussian random variable specified using weighted sigma points, which completely capture true mean and covariance of the distribution. Another solution for the general filtering problem is to use sequential Monte Carlo methods. It is particle filtering PF based on sequential importance sampling where the samples (particles) and their weights are drawn from the posterior distribution.