| For a long time, ocean environment monitoring has been faced the contradiction between resolution and working distance, time-space scale and equipment resources. The ocean dynamic model has a wide monitoring range, but the resolution scale and accuracy are limited. Acoustic inversion depends on local observation nodes, but it is sensitive to small and medium scale dynamic processes. In order to obtain a more accurate method of ocean information acquisition, the coupling relationship between the ocean dynamic processes and acoustic propogation is fully excavated and the acoustic inversion is combined with the dynamic model calculation.Considering the inversion of environment parameters in watercolumn requires the bottom geoacoustic information, the matched field inversion is employed to estimate the geoacoustic parameters. Based on the sensitivity analysis of the objective function, the genetic algorithm which is an efficient algorithm for global search is used in matched field inversion. For the acoustic measurement system of sound speed profiles inversion, the thesis makes a comparative analysis of the single frequency and wide band signal for the geoacoustic inversion using a vertical line array. Besides, a preliminary study is performed on the geoacoustic parameter inversion using distributed sensor nodes.Considering the spatial-temporal evolution characteristics of sound speed profiles, a Weighted Ensemble Kalman filter (WEnKF) based on the Ensemble Kalman filter (EnKF) is proposed, which takes the background set sample distribution into account, and the sample weights are added to alleviate the Gauss approximation hypothesis. The results of the experiment data processing show that the WEnKF outperforms the EnKF, especially in the case of fast changing SSPs. Meanwhile, the state evolution equation is established by using the ROMS forecast. With the combination of the acoustic pressure observation equation and the state equation, the state-space model is obtained with thecoupling of the sound field and the dynamic processes. The examples show that the WEnKF with the ROMS forecast can effectively improve the accuracy of the range-dependent inversion.Finally, a software kit is developed, which includes the inversion of sound speed profiles in water column with the WEnKF and the geoacoustic inversion in sediment. The overall architecture and functional modules of the software are introduced in the thesis. |
