强非线性时间演化声速剖面的序贯反演

受海面波浪起伏、降雨和内波等海洋动力学过程的影响,浅水声速剖面的时间演化具有高度非线性,针对该问题提出使用改进的粒子滤波方法进行声速剖面序贯反演.该方法通过建立声速剖面的经验正交模型(EOF)以及描述声速剖面时间演化特征的状态空间模型,将声速剖面反演问题建模为状态跟踪问题,利用不敏粒子滤波(UPF:Uncented Particle Filter)算法进行声速剖面序贯反演。仿真试验通过实测声速剖面数据和先验地声参数信息产生接收声场数据,再利用模拟声场数据估计声速剖面的时间变化.结果表明,相比于集合卡尔曼滤波(EnKF:Ensemble Kalman Filter),在计算效率等同的情形下,该方法可以在状态参数的时间跳变点保持良好的跟踪性能,一定程度上克服了现有反演算法在跳变点发散的问题,可以有效提高声速剖面反演精度,尤其在声速剖面时变性较强时具有显著优势.

Sequential inversion of highly nonlinear time-evolving sound speed profiles

Affected by ocean dynamic processes such as sea waves,rainfall and internal waves,the time evolution of sound speed profiles(SSPs) in shallow water is highly nonlinear.To solve the problem,an algorithm of the improved particle filter is implemented for the tracking of time-evolving SSPs.Based on the Empirical Orthogonal Functions(EOFs) and the state-space model which describe the evolution characteristics of SSPs,sequential inversion of SSPs are carried out through the acoustic pressure data received by the Vertical Line Array(VLA) using UPF.Time-evolving SSPs are estimated via the acoustic array data simulated by the measured SSPs and prior seabed acoustic properties.The algorithm was validated and result shows that under the comparable computational efficiency,the UPF-based method can overcome the divergences of Ensemble Kalman Filter(EnKF) algorithm and keeps up perfect tracking performance at the jump time.The estimated accuracy can be effectively increased,especially in the case of strongly time-evolving SSPs.