Extended Kalman filter-based data fusion for lnertial-Doppler AUV navigation

摘要

Use of Auronomous Underwater Vehicles(AUVs) has been gaining momentum in both military and commercial applications.While some enabling technologies such as autoaomy of maneuvering and power efficiency have growninto mature, navigation remains to be a limiting factor to make AUV truly autonomous.Both commonly-used navigation devices, Inertial Navigation system (INS) and Doppler Veiocity Log (DVL), have obvious weaknesses. A practical approach widely used for long travel distance AUVs is based on Inertial-Doppler combinational navigation and Kalman Filter has been frequently applied to fuse the navigation information from both devices.However the AUVdynamics is often govemed by a nonlinear relationship.Itis thus necessary to replace the regular Kalman filter with the Extended Kalman Filter (EKF), which is able to handle nonlinear recursion. This papes presents a general framework to apply the EKF to Inertial-Doppler navigation based on a general dynamic model for AUV Cruise. Numeric simulations are implemented to verify the effectiveness of the developed approach.