Generation of Vessel Track Characteristics Using a Conditional Generative Adversarial Network (CGAN)

ABSTRACT

Machine learning (ML) models often require large volumes of data to learn a given task. However, access and existence of training data can be difficult to acquire due to privacy laws and availability. A solution is to generate synthetic data that represents the real data. In the maritime environment, the ability to generate realistic vessel positional data is important for the development of ML models in ocean areas with scarce amounts of data, such as the Arctic, or for generating an abundance of anomalous or unique events needed for training detection models. This research explores the use of conditional generative adversarial networks (CGAN) to generate vessel displacement tracks over a 24-hour period in a constraint-free environment. The model is trained using Automatic Identification System (AIS) data that contains vessel tracking information. The results show that the CGAN is able to generate vessel displacement tracks for two different vessel types, cargo ships and pleasure crafts, for three months of the year (May, July, and September). To evaluate the usability of the generated data and robustness of the CGAN model, three ML vessel classification models using displacement track data are developed using generated data and tested with real data.

Data Availability Statement

The AIS data set from 2022 used in this experiment is openly available at https://coast.noaa.gov/htdata/CMSP/AISDataHandler/2022/index.html. It was acquired from the National Oceanic and Atmospheric Administration (NOAA) Vessel Traffic Data repository https://marinecadastre.gov/AIS/.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Abbreviations

Artificial Intelligence	=	AI
Automatic Identification System	=	AIS
Course Over Ground	=	COG
Conditional Adversarial Network	=	CGAN
Discriminator	=	$\mathcal{D}$
Generative Adversarial Network	=	GAN
Generator	=	$\mathcal{G}$
Kolmogorov-Smirnov	=	KS
Machine Learning	=	ML
Multi-Layer Perceptron	=	MLP
Maritime Mobile Service Identity	=	MMSI
Rectified Linear Unit	=	ReLU
Speed Over Ground	=	SOG

Supplementary Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08839514.2024.2360283

Notes

1. Annotations are labels that assign an instance to a category or group.

2. AIS data provides details about vessel motion including speed, location coordinates, course, etc.

3. The AIS provides both dynamic and static information with respect to vessels. This includes information related to: time, date, speed, course, heading, ship type, ship dimensions, etc.

4. Web site: https://marinecadastre.gov/AIS/data

5. Skew refers to the distortion of symmetrical distribution within the data.

6. Adaptive Moment Estimation is the algorithm used to help determine the optimal network parameters.

7. The batch size is a number of instances that are processed before updating the weights.

8. An epoch is a full pass through the entire training set.

9. This statistic has a range of 0 to 1.

10. The F1 score is the harmonic mean of the precision and recall. The desired value is one which indicates perfect precision and recall.