The refined resilient model for underwater acoustic positioning

Abstract

The variation of sound speed is a major factor affecting the accuracy of the acoustic positioning underwater. A refined resilient model is constructed which includes temporal variations and horizontal variations on the basis of the simple resilient observation model which includes range bias and time bias parameters. The integrated estimation method and the stepwise estimation method were used to calculate the temporal and spatial variations of the sound speed using the GNSS-Acoustic (GNSS-A) measurements. The sound speed model was used in the long baseline (LBL) underwater acoustic positioning, and the position of the transducer calculated by acoustic ranges is compared with that provided by GNSS. The results show that, there are obvious temporal and spatial variations in the sound speed in different water area. The temporal variation of the sound speed is less than 0.8 m/s and the spatial variation is less than 0.15 m/s. The variation of sound speed under water not only shows obvious periodicity, but also presents negative gradient in the horizontal direction. When the coordinates of the transducer are calculated by using the sound speed model considering temporal and spatial variations constructed by the integrated estimation, the precision of the coordinates is improved by about 28%, compared with that calculated using the constant sound speed model. The 3D root mean square (RMS) of the LBL underwater acoustic positioning is about 0.73 m, and the RMS is less than 0.13 m in the vertical direction, when the coordinates of the ship bottom transducer are calculated within the area covered by five seafloor transponders. It indicates that the coordinate accuracy of the five seafloor transponders is high in the vertical direction.