Research on the navigation system of a class of underwater vehicle based on USBL

This paper presents a novel unscented Kalman filter (UKF) used for navigation of Human Occupied Vehicle (HOV) based directly on the nonlinear sensor readings of an Ultra-short Baseline (USBL), a Doppler Velocity Log (DVL), a fiber optic gyrometer and a depth sensor. The HOV motion and the USBL observations are highly non-linear processes which contain unknown noise. A UKF is therefore chosen as a suitable data fusion technique. For the low rate positional measurements of USBL and the drift error of the DVL, the presented UKF fuses the information from these sensors to produce a more accurate estimate of three-dimensional position, orientation (heading), and velocity of the HOV. MATLAB simulations conducted with respect to the data obtained from previous sea trial illustrate the effectiveness of our proposed method.