Extra colloquium Computer Science Dr. A. Abdi, University of Twente

Title: Deep Vessel: A Deep Learning Model for Vessel Arrival Time Prediction

Abstract:

Maritime transportation is a major mode of transportation for domestic and international trade due to its large capacity and coverage of a huge amount of world trade. The logistic community of shippers have however struggled to predict the precise arrival of the world’s ocean or sea vessels with reliable certainty. Therefore, this study explores the use of a new deep learning-based method to predict vessel arrival time that could eventually be incorporated into a business intelligence dashboard (called RATP). To the best of our knowledge, a deep learning-based method in which a unified feature set which is representative of Automatic Identification System (AIS) dataset and augmented information, has not been thoroughly studied for the prediction of the arrival time of vessels. On the other hand, the existing methods fail to provide desirable results due to a shallow neural network architecture along with hand-engineered features. Thus, we present a novel deep-learning-based method consisting of stacked LSTMs. The stack of Recurrent Neural Network (RNN) layers are composed with Long Short-Term Memory (LSTM) to take advantage of sequential processing and the longitudinal nature of AIS, meteorological and vessel information. To verify the effectiveness of RATP, we conduct our experiment on large-scale datasets. The encouraging results show that the RATP achieves significant accuracy. Furthermore, these positive results demonstrate that i) feature vectors such as AIS and augmented information can improve the accuracy of vessel arrival time prediction; ii) our method learns from this unified feature set and can obtain significant performance compared to one that learns from a subset of the features.