Novel Initialization Method for Nonlinear Kalman Filters
Yu Zhe Wang, Hang Su, Yu Kevin Cao, Han Liu · Advanced materials research · 2013
A novel initialization method for nonlinear Kalman filter was proposed aimed at the defects of the ordinary initialization method. This method was stricter than the ordinary method in theory and could guarantee scientific compare between nonlinear Kalman filters based on Gaussian assumptions. Simulation results showed that the novel method greatly reduced the impact of initialization and reflected more the estimation error due to the type of nonlinear Kalman filters rather than initialization.