An adaptive stratification algorithm based on gradient fitting deviation and its application to acoustic ray-tracing algorithm

Abstract

Accurate underwater positioning is a challenging task for complex marine environments where the study of sound velocity profiles (SVPs) is essential. To overcome the problem that the original SVP reduces the computational efficiency of the ray-tracing algorithm due to great capacity of data, this paper simplifies the SVP. The nature of the SVP and its influence on accurate underwater positioning are discussed. An adaptive stratification algorithm applicable to the ray-tracing algorithm is proposed, based on the gradient fitting deviation and characterized by the fitting treatment of the sound velocity gradient by the Fourier method. Measured SVP data and acoustic positioning data were used to validate this algorithm in the context of acoustic ray-tracing algorithm, to analyze and compare the simplification rates before and after SVP simplification, and the effect on the positioning accuracy. The experimental results in the shallow sea (water depth of 60 m) show that the simplification rate of the SVP is more than 95.56%, and the RMS of the deviation of the ray-tracing algorithm is better than 0.0112 m when determining the optimal number of unfolding levels and threshold; the experimental results in the deep sea (water depth of 1900 m) show that the simplification rate of the SVP is more than 99%, and the RMS of the deviation of the ray-tracing algorithm is better than 0.5 m. This indicates that, the algorithm can reconstruct the original SVP well and implement the automatic hierarchical simplification of the SVP while ensuring the positioning accuracy.