•  
  •  
 

Abstract

The underwater glider has the advantages of low power, long endurance and high accuracy. Micro-Electro-MechanicalSystem (MEMS) grade inertial sensors are more suitable for an underwater glider because of their low cost and small size. Models of MEMS sensor noises which include not only the white noises and random walk terms but also the bias instabilities of the sensor noises are analyzed. The integrated navigation system for the underwater glider is composed of a dead reckoning (DR) module, an inertial navigation system (INS) based on MEMS sensors aided by a tri-axis magnetic sensor. Due to the inherent error characteristics, MEMS grade devices suffer from the non-stationary stochastic sensor errors and nonlinear inertial errors which cannot be well handled by the conventional filter algorithms, this paper proposes extended Kalman filter (EKF) fusing Runge-Kutta (RK4) algorithm (EKF/RK4) which can implement the data fusion of multisensor. The proposed EKF/RK4 can take advantage of the EKF to achieve the optimal estimation of attitude and position and then make better use of the RK4 to further improve the estimation accuracy. In order to evaluate the effectiveness of the proposed algorithm, the EKF/RK4 algorithm is applied to the underwater navigation system designed in our lab and a series of land experiments are performed. The performance of the proposed EKF/RK4 algorithm based on our navigation system is analyzed and compared with the traditional algorithms. The experiment results show that the proposed algorithm is more effective in reducing the attitude and position errors than KF/RK4 and EKF.

Download
COinS