<i>GeoTrackNet</i>—A Maritime Anomaly Detector Using Probabilistic Neural Network Representation of AIS Tracks and <i>A Contrario</i> Detection

Nguyen, Duong; Vadaine, Rodolphe; Hajduch, Guillaume; Garello, R.; Fablet, Ronan

doi:10.1109/tits.2021.3055614

Cited by 61 publications

(49 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Yet, some studies argued that such representation might not be relevant for neural-network based methods since movement trajectories inherently contain 2D spatial information which might not be easily perceivable in simple time-series. For instance, previous works investigated one-hot representations initially introduced for text data [36] as well as trajectory image [14]. Here, we explored distance matrix representations as the distance matrix of positions has been shown to capture relevant spatial information for protein structure prediction [46,60].…”

Section: Resultsmentioning

confidence: 99%

Deep inference of seabird dives from GPS-only records: performance and generalization properties

Roy

Fablet

Lanco-Bertrand

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Seabirds are considered as suitable indicators for the study of marine ecosystems, since their foraging strategies provide a real-time response to complex ecosystem dynamics. By deploying GPS sensors on seabirds it is possible to obtain their trajectories, and deep learning have recently shown promising results for the classification of animal behaviour from trajectory data. Yet there is still lot of investigation needed in terms of network architectures, data representation but also to demonstrate the generalization properties of these approaches. From a database of about 250 foraging trajectories derived from GPS data deployed simultaneously with pressure sensors for the identification of dives, this work has consisted in training deep networks in a supervised manner for the prediction of dives from trajectory data. In this study, we confirm that deep learning allows better dive prediction than usual methods such as Hidden Markov Models for two distinct seabirds species. We propose a novel deep learning model for trajectory data. It combines the computational efficiency of convolutional neural networks to distance-matrix-based representations of trajectory data. Our model considerably increases the ability of deep networks to infer behaviour, as well as their stability to different data inputs. The considered trajectory data representation might enable deep networks to better capture spatial information than from longitude and latitude time-series considered in previous works.

show abstract

Section: Resultsmentioning

confidence: 99%

Deep inference of seabird dives from GPS-only records: performance and generalization properties

Roy

Fablet

Lanco-Bertrand

2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…This is a general limitation of clustering-based approaches, which may explain why they are yet used in operational systems, to the best of our knowledge. As illustrated in [22], [26], it is not always possible to assign a given AIS trajectory to a predefined cluster associated with a maritime route. Besides, the clustering may also result in small clusters to best represent the variability of the AIS trajectories.…”

Section: Related Workmentioning

confidence: 99%

“…However, as discussed in [22] and [26], it is difficult to encode the complex and multimodal movement patterns of vessels in this feature space. A natural idea is to expand to a higher dimensional space, i.e.…”

Section: B Discrete and Sparse Representation Of Ais Datamentioning

confidence: 99%

TrAISformer-A generative transformer for AIS trajectory prediction

Nguyen¹,

Fablet²

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Modelling trajectory in general, and vessel trajectory in particular, is a difficult task because of the multimodal and complex nature of motion data. In this paper, we present TrAISformer-a novel deep learning architecture that can forecast vessel positions using AIS (Automatic Identification System) observations. We address the multimodality by introducing a discrete representation of AIS data and re-frame the prediction, which is originally a regression problem, as a classification problem. The model encodes complex movement patterns in AIS data in high-dimensional vectors, then applies a transformer to extract useful long-term correlations from sequences of those embeddings to sample future vessel positions. Experimental results on real, public AIS data demonstrate that TrAISformer significantly outperforms state-of-the-art methods.

show abstract

“…ML (or deep learning, DL, as being branded lately) algorithms, such as LSTM and LSTM encoder decoder [38], [39], have been shown to be promising in sequence-tosequence or time-series prediction in a number of applications, including trajectory prediction tasks [17], [31], [40], [41]. In addition, [16] applied the Variational RNN to learn the probabilistic distribution of trajectories, which help them to detect anomalous trajectories.…”

Section: Related Workmentioning

confidence: 99%

“…Evaluation Metrics: In the case of anomalous trajectory detection, there is no reference dataset for anomalous tracks [16], a quantitative analysis in terms of accuracy or a false alarm rate is therefore not feasible. In contrast, we can evaluate the EDL classifiers for detecting OOS and UT anomalies using the following standard metrics: the accuracy measure, confusion matrix, false positive or negative rate.…”

Section: A Experimental Setupmentioning

confidence: 99%

Leveraging Graph and Deep Learning Uncertainties to Detect Anomalous Trajectories

Singh¹,

Fowdur²,

Gawlikowski³

et al. 2021

Preprint

View full text Add to dashboard Cite

Understanding and representing traffic patterns are key to detecting anomalies in the maritime domain. To this end, we propose a novel graph-based traffic representation and association scheme to cluster trajectories of vessels using automatic identification system (AIS) data. We utilize the (un)clustered data to train a recurrent neural network (RNN)-based evidential regression model, which can predict a vessel's trajectory at future timesteps with its corresponding prediction uncertainty. This paper proposes the usage of a deep learning (DL)-based uncertainty estimation in detecting maritime anomalies, such as unusual vessel maneuvering. Furthermore, we utilize the evidential deep learning classifiers to detect unusual turns of vessels and the loss of AIS signal using predicted class probabilities with associated uncertainties. Our experimental results suggest that using graphbased clustered data improves the ability of the DL models to learn the temporal-spatial correlation of data and associated uncertainties. Using different AIS datasets and experiments, we demonstrate that the estimated prediction uncertainty yields fundamental information for the detection of traffic anomalies in the maritime and, possibly in other domains.

show abstract

GeoTrackNet—A Maritime Anomaly Detector Using Probabilistic Neural Network Representation of AIS Tracks and A Contrario Detection

Cited by 61 publications

References 41 publications

Deep inference of seabird dives from GPS-only records: performance and generalization properties

Deep inference of seabird dives from GPS-only records: performance and generalization properties

TrAISformer-A generative transformer for AIS trajectory prediction

Leveraging Graph and Deep Learning Uncertainties to Detect Anomalous Trajectories

Contact Info

Product

Resources

About