Forward and Inverse Stochastic Filtering for Inertial Sensor Calibration

J. Fox (Germany)


sigma-point Kalman filter, intertial navigation, calibration


Extended Kalman filters have long been applied to sen sor fusion in navigation tasks. They can be used to es timate both the states and the parameters of the dynamic system. In recent years, so-called sigma-point Kalman fil ters with an improved estimation accuracy compared to ex tended Kalman filters have emerged. This work shows how these filters can be applied to calibrate an inertial measure ment unit used for unaided navigation. Two different filter structures are proposed: a forward filter models the whole navigation process (states and parameters) while an inverse filter performs only a parameter estimation.

Important Links:

Go Back