Artificial Intelligence

High-Performance Technologies for Autonomous Port Operations

COLOSSUS incorporates state-of-the-art technologies designed to deliver high on-site performance for connected unmanned vehicles (UVs), ensuring robustness, interoperability, and operational efficiency in demanding port and maritime environments.

Multi-modal Data Fusion and Analysis

COLOSSUS adopts state-of-the-art multi-modal data-efficient fusion techniques to handle the big volume of diverse data from a range of sensors on the UVs and 3D structure models and to ensure data quality and representation. An AI-based analysis module will further improve the quality of the data and perform initial identification of incidents, optimising real-time monitoring while minimising data redundancy.

3D Reconstruction and modelling

An advanced system to create high-resolution, real-time models of underwater structures for precise structural analysis and proactive maintenance. A lightweight version will be explored to be deployed on the USV, to have direct access to the sonar and LiDAR sensors, while also reducing the amount of data transmitted back to base.

Intelligent Swarm Navigation

An explainable ML framework based on swarm intelligence will be proposed and demonstrated for trajectory planning and swarm formation control of diverse types of autonomous unmanned vehicles. Optimised unmanned patrols will be scheduled to cover the vast port multi-terrain aera, whereas in case of incident detection, and depending on the conditions surrounding it, this framework will coordinate the available and appropriate UVs (e.g., in an intrusion incident along the water perimeter only the USVs and UAVs will be deployed) to further explore and potentially address the incident.

Skip to content