| Abstract [eng] |
Modern deep learning algorithms are able to handle large amounts of data and therefore are particularly important in automating vessel movement prediction in intensive shipping. This would be one of the support tools for monitoring, managing the increasing maritime traffic and its participants. In the work, the deep recurrent network is realized by applying Java technology and integrating into it one of the most popular library of the neural networks Deep Learning 4 Java. Applying deep learning algorithm, a recurrent network (LSTM) is created that is able to predict the further vessel movement. The developed architectural model is based on sequences when data change over time, therefore the work investigates the most optimal data structure and network hyperparameters, which aim to obtain the most accurate prediction results. Different recurrent network architectures were used to compare the results: basic LSTM, LSTM stack, autoencoder (AE) and variational autoencoder (VAE). During the study found that the most accurate prediction is performed with an autoencoder architecture. The study was performed on a specific sample of data from the Dutch coastal region and the proposed algorithm can be applied as one of the ways to improve maritime safety. |
