在海洋中存在浅水声道、深水声道和会聚区，在这些区域声波以声线的形式保持一定能量上下波动传播得很远。对于测距类长基线水声定位系统, 一般基于声线近似单调曲线传播或近似直线传播的假设。考虑声线在声道中波动传播，需要发展新的水下声学定位方法。本申请基于声速剖面已知的条件下，研究声线随时间变化的声线方程，将海洋中的弯曲声线分解为基于时间积分的水平距离定位方程和水深约束定位方程，采用最小二乘法和抗差最小二乘法估计水下目标位置，克服声线上下波动带来的斜距误差大问题，并且定位模型不受距离的限制；基于时间积分的计算声线水深，声线向下传播水深增量为正，向上为负，避免声线在声道中折射或反射波动前进的模糊度问题；将不同初始掠射角条件下以时间积分计算声线族，将声线族末端的目标水深作为约束条件计算初始掠射角和目标水深的近似值，避免近似值搜索求解问题。所研究的方法同适用于短基线和超短基线测距误差改正。

There are surface sound channel, underwater sound channel and convergence area in the ocean, in these areas, the sound wave can fluctuate far away with some energy in the form of sound rays. The ranging-like long-baseline underwater acoustic positioning systems is generally based on that sound rays are approximately straight line or monotone curve. Considering the wave propagation of sound rays in the ocean sound channel, it is necessary to develop a novel approach of underwater acoustic localization. Based on the known sound speed profile, the bending sound rays in the ocean are decomposed into horizontal distance localizationequation and water depth localization equation on time integral, and least square method and robust least square method are adopted to estimate of localizationof underwater target, which overcomes the problem of large pitch error caused by the fluctuation of sound rays, and localization model is not affected by distance. The depth of sound rays are calculated based on time integral, which increment is positive downward and negative upward, which avoids the ambiguity of the refraction or reflection of sound rays in the ocean sound channel. Sound rays family is calculated by time integral under different initial grazing angles, and the approximate values of the initial grazing angle and the target water depth are calculated with the target water depth at the end of the sound ray family as the constraint condition, which avoids the problem of search of approximate values. The method is also suitable for range-like short baseline and ultra-short baseline location.