Iterated square root unscented Kalman filter and its application in deep sea vehicle navigation

Abstract

It is of vital importance to develop a high accuracy and fast convergence algorithm in deep sea navigation system, since location is essential for the sake of scientific survey and safety in very hazard environment. Unscented Kalman Filter (UKF) is the type of filter, which is designed in order to overrun this problem. However, in case of state estimation of the Human Occupied Vehicle (HOV) via the sensor data obtained from a Long Baseline (LBL) acoustic positioning system, a Doppler Velocity Log (DVL), a depth sensor and a motion sensor, where the nonlinearity degree of dynamic model is high and the operating environment is complex, UKF may give inaccurate results. In this study an iterated square root Unscented Kalman Filter (ISRUKF) is presented. An iterated measurement update procedure is included to increase the approximation accuracy of nonlinear state estimates, and a square root version of UKF is conducive to guarantee numerical stability of the algorithm. Compared with the UKF and square root Unscented Kalman Filter (SRUKF), used in deep-sea vehicle navigation system, the ISRUKF algorithm has potential advantages in convergence speed and location accuracy. Extensive experiment researches have been conducted by using the data obtained from previous sea trial to demonstrate its superiority.