Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Predictive Modeling of Maritime Radar Data Using Transformer Architecture

Bjorna Qesaraku, Jan Steckel

TL;DR

This survey identifies a critical gap: while transformer models have advanced AIS trajectory and sonar frame prediction, their application to maritime radar frame prediction remains unexplored. It argues that frame-level radar forecasting preserves rich scene context, supports anomaly detection, and complements trajectory-based methods for robust, anticipatory perception. By reviewing traditional, ML, and DL approaches, as well as transformer-enabled work in related domains and the MOANA dataset, the paper outlines concrete future directions, including adapting EchoPT to radar, self-supervised pretraining, and efficient on-board implementations. The work underscores the practical significance of radar-frame forecasting for collision avoidance and degraded-sensing resilience in autonomous maritime operations.

Abstract

Maritime autonomous systems require robust predictive capabilities to anticipate vessel motion and environmental dynamics. While transformer architectures have revolutionized AIS-based trajectory prediction and demonstrated feasibility for sonar frame forecasting, their application to maritime radar frame prediction remains unexplored, creating a critical gap given radar's all-weather reliability for navigation. This survey systematically reviews predictive modeling approaches relevant to maritime radar, with emphasis on transformer architectures for spatiotemporal sequence forecasting, where existing representative methods are analyzed according to data type, architecture, and prediction horizon. Our review shows that, while the literature has demonstrated transformer-based frame prediction for sonar sensing, no prior work addresses transformer-based maritime radar frame prediction, thereby defining a clear research gap and motivating a concrete research direction for future work in this area.

Predictive Modeling of Maritime Radar Data Using Transformer Architecture

TL;DR

This survey identifies a critical gap: while transformer models have advanced AIS trajectory and sonar frame prediction, their application to maritime radar frame prediction remains unexplored. It argues that frame-level radar forecasting preserves rich scene context, supports anomaly detection, and complements trajectory-based methods for robust, anticipatory perception. By reviewing traditional, ML, and DL approaches, as well as transformer-enabled work in related domains and the MOANA dataset, the paper outlines concrete future directions, including adapting EchoPT to radar, self-supervised pretraining, and efficient on-board implementations. The work underscores the practical significance of radar-frame forecasting for collision avoidance and degraded-sensing resilience in autonomous maritime operations.

Abstract

Maritime autonomous systems require robust predictive capabilities to anticipate vessel motion and environmental dynamics. While transformer architectures have revolutionized AIS-based trajectory prediction and demonstrated feasibility for sonar frame forecasting, their application to maritime radar frame prediction remains unexplored, creating a critical gap given radar's all-weather reliability for navigation. This survey systematically reviews predictive modeling approaches relevant to maritime radar, with emphasis on transformer architectures for spatiotemporal sequence forecasting, where existing representative methods are analyzed according to data type, architecture, and prediction horizon. Our review shows that, while the literature has demonstrated transformer-based frame prediction for sonar sensing, no prior work addresses transformer-based maritime radar frame prediction, thereby defining a clear research gap and motivating a concrete research direction for future work in this area.

Paper Structure

This paper contains 19 sections, 1 equation, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Background
  3. Maritime Radar Fundamentals
  4. Transformer Architectures for Prediction
  5. Advantages of Predicting Radar Frames
  6. Literature Review
  7. Traditional and Machine Learning Approaches in Maritime Applications
  8. Traditional Methods
  9. Machine Learning (ML) Methods
  10. Deep Learning (DL) Approaches in Maritime Applications
  11. CNN for Target Detection
  12. RNNs for Trajectory Prediction
  13. Transformer-Based Approaches for Next-Frame Prediction
  14. AIS Trajectory Prediction
  15. Object Detection with Transformers
  16. ...and 4 more sections

Figures (2)

  • Figure 1: Annotated maritime scene showing multiple objects (small vessels, distant ships and navigation buoys and flags) that must be detected and classified by the perception system. Base photograph by Charlotte Clark.
  • Figure 2: Comparison of AIS and radar observations. (a) AIS display/control unit shows only cooperative vessels that can broadcast AIS messages like identity, voyage data, etc. Image by Clipper, "Ais dcu bridge", 2006, Wikimedia Commons, licensed under CC BY 2.5 AIS2006. (b) Example of an X-band radar frame from MOANA dataset Jang2024 containing the full radar scene, including shoreline returns and additional small echoes that may correspond to clutter, navigation aids, or buoys.