Xiaoyu Zhu

Geoacoustic inversion is important for seabed geotechnical applications. It can be formulated as a problem that seeks an optimal solution in a high-dimensional parameter space. The conventional inversion approach exploits optimization methods with a pre-defined search strategy whose hyperparameters need to be fine-tuned for a specific scenario. A framework based on the deep-Q network is proposed in this paper and the environment and agent configurations of the framework are specially defined for geoacoustic inversion. Unlike a conventional optimization method with a pre-defined search strategy, the proposed framework determines a flexible strategy by trial and error. The proposed framework is evaluated by two case studies for estimating the shear wave velocity profile. Its performance is compared with three global optimization methods commonly used in underwater geoacoustic inversion. The results demonstrate that the proposed framework performs the inversion more efficiently and accurately.

We propose a time-frequency fused underwater acoustic source localization method based on self-supervised learning with contrastive predictive coding. Firstly, two feature extractors are trained to solve the pretext task (predicting the future) based on the unlabeled acoustic signals in the time and frequency domains, respectively. Next, encoders with frozen parameters are taken from the trained feature extractors for extracting the high-level features in the time and frequency domains. During the training stage of the source localizer, features extracted by two encoders are concatenated together as a time-frequency fused feature vector and fed into a 3-layer multi-layer perceptron for solving the downstream task (source localization) based on a tiny labeled dataset. This method is assessed on the SWellEx-96 Experiment and compared with several alternative methods. The performance analysis confirms the promising performance of our proposed method.

Underwater source localization is an important task, especially for real-time operation. Recently, machine learning methods have been combined with supervised learning schemes. This opens new possibilities for underwater source localization. However, in many real scenarios, the number of labeled datasets is insufficient for purely supervised learning, and the training time of a deep neural network can be huge. To mitigate the problem related to the low number of labeled datasets available, we propose a two-step framework for underwater source localization based on the semi-supervised learning scheme. The first step utilizes a convolutional autoencoder to extract the latent features from the whole available dataset. The second step performs source localization via an encoder multi-layer perceptron trained on a limited labeled portion of the dataset. To reduce the training time, an interpretable feature selection (FS) method based on principal component regression is proposed, which can extract important features for underwater source localization by only introducing the source location without other prior information. The proposed approach is validated on the public dataset SWellEx-96 Event S5. The results show that the framework has appealing accuracy and robustness on the unseen data, especially when the number of data used to train gradually decreases. After FS, not only the training stage has a 95% acceleration but the performance of the framework becomes more robust on the receiver-depth selection and more accurate when the number of labeled data used to train is extremely limited.

In this paper, we propose an interpretable feature selection method based on principal component analysis (PCA) and principal component regression (PCR), which can extract important features for underwater source localization by only introducing the source location without other prior information. This feature selection method is combined with a two-step framework for underwater source localization based on the semi-supervised learning scheme. In the framework, the first step utilizes a convolutional autoencoder to extract the latent features from the whole available dataset. The second step performs source localization via an encoder multi-layer perceptron (MLP) trained on a limited labeled portion of the dataset. The proposed approach has been validated on the public dataset SwllEx-96 Event S5. The result shows the framework has appealing accuracy and robustness on the unseen data, especially when the number of data used to train gradually decreases. After feature selection, not only the training stage has a 95\% acceleration but the performance of the framework becomes more robust on the depth and more accurate when the number of labeled data used to train is extremely limited.

An evaporation duct is a particular atmospheric duct that is crucial for marine vessel communication, and one of the most significant indexes to assess it is evaporation duct height (EDH). In this letter, we propose a new method based on a multilayer perceptron (MLP), a classic network in deep learning, to predict the EDH. After so many experiments, the structure of MLP for EDH estimation is chosen as five hidden layer with rectified liner unit activation function. Since this method is essentially some kind of regression, the mean-squared error is chosen as the loss function. To accelerate the training convergence, we choose a self-adaptive optimization scheme called adaptive moment estimation. There are some field observation data sets obtained from some sea areas in northern hemisphere, by which the MLP is trained, to predict the EDH. For single-regional prediction, we use the same input data sets with the Paulus-Jeske (P-J) model, one of the ideal operational models, to predict the EDH. In comparison with the P-J model, the prediction accuracy using our method is significantly escalated in all experimental sea areas, which reveals the efficiency of our method. By cross prediction in distinct sea areas, the consistency between the method and the theory is verified, and this letter yields a new approach of evaporation duct prediction.