Deep learning-based prediction of ship transit time

Vessel traffic system (VTS) operators instruct ships to wait for entry and departure to sail one-way in order to prevent ship collision accidents in harbors with narrow routes. At present, these instructions are not based on scientific or statistical data. Consequently, there was a significant deviation depending on the individual capabilities of the VTS operators. Accordingly, in this study, a 1D convolutional neural network model was built by collecting ship and weather data to predict the exact travel time for ship arrival/departure waiting for instructions at the harbor. The proposed deep learning model was confirmed to be improved by more than 5.9% compared to other ensemble machine learning models. Through this study, it is possible to predict the time required to enter and depart a vessel in various situations; therefore, the VTS operators are expected to assist in providing accurate information to the vessel and determining the waiting order. •In this study, important features, such as ship heading posture at berthing, which affect transit time, were extracted.•The algorithms for calculating the transit time of inbound and outbound ships were designed.•This study built a 1D convolutional neural network model to predict the exact transit time.