This paper proposes a simplified algorithm for reducing the computational load of the conventional underwater integrated navigation system. The system usually comprises a three-dimensional accelerometer, a three-dimensional gyroscope, a three-dimensional Doppler Velocity Log (DVL) and a data fusion algorithm, such as a Kalman Filter (KF). Since the expected variations of roll, pitch and depth are small, these quantities are assumed to be constant, and the proposed system is designed in a two-dimensional form. Due to the low speed of the vehicle, the nonlinear dynamic equation of the velocity can be simplified in a linear form. We also simplify the conventional KF in order to avoid matrix multiplications and matrix inversions. The performance of the designed system is evaluated in a sea trial by an Autonomous Underwater Vehicle (AUV). The results show that the proposed system can significantly reduce the computational load of the conventional integrated navigation system without a significant reduction in position and velocity accuracy.

No abstract

Strapdown Inertial Navigation System (SINS) estimates position, velocity, and attitude of vehicle using the signals measured by accelerometer and gyroscope and is based on dead-reckoning principle. Due to different imperfections in measurements, and the consecutive integration of the acceleration signals, estimation error increases with time and it is acceptable only for short times in the low cost SINS. In order to reduce the error, auxiliary sensors together with inertial sensors are utilized and to combine the data estimated by SINS with the signals measured by auxiliary sensors, data fusion methods based on direct and indirect Kalman filtering is used. The feedforward and feedback structures are two common approaches of the indirect filtering. In this paper, an underwater integrated navigation system has been designed using indirect filtering approaches due to lower computational load and higher reliability with respect to the direct approach. The auxiliary sensors used consist of DVL, Gyrocompass, and depthmeter. The performance of the designed system has been studied using real measurements. The experimental results showed that the root mean square error in the estimated position for feedforward structure is reduced from 3.2 to 0.2 percent of the travelled distance when using feedback structure.

No abstract

No abstract