Maritime Innovation
Robust Bathymetric Surveys with Flow-Aware Reinforcement Learning
High-resolution underwater depth data is vital for navigation safety, hydrological studies, and infrastructure assessment. These surveys are often conducted in areas subject to strong currents, where undetected scour can severely undermine the structural integrity of transport infrastructure. Other high-flow scenarios requiring regular inspection include dam water intakes and the monitoring of industrial wastewater discharges. In such conditions, strong currents can significantly disrupt observer trajectories and reduce the accuracy of bathymetric measurements.
Given the operational hazards associated with high-flow environments, advanced control methods are essential for dependable navigation and surveillance. Intractābilis has developed a pioneering reinforcement learning-based control framework that directly integrates local flow field measurements. This crucial input, traditionally absent from conventional systems, transforms previously intractable navigation challenges into manageable tasks, markedly enhancing operational performance.
Our approach further incorporates spatial attention mechanisms for sensor fusion and temporal attention for modelling environmental dynamics, implemented within a novel Transformer-based architecture. This combination not only improves mission success rates and route efficiency compared with classical and learning-based baselines, but also demonstrates robust generalisation across reinforcement learning frameworks and varying obstacle densities.
Quantum Annealing for Maritime Shipment Rerouting
In a maritime transport network comprising hubs and various transportation assets operating on defined routes, a series of transport requests must be fulfilled. Each request specifies a quantity of goods to be moved from an origin to a designated destination. The overarching goal is to ensure these deliveries are completed at the lowest possible cost, while meeting all logistical requirements.
For the specific case under consideration, the total cost of transporting all goods can be minimised by utilising a single container for the entire shipment. The optimal resolution to the maritime shipment rerouting challenge involves determining a route that visits all relevant hubs while achieving the shortest overall travel distance.
Intractābilis has developed a quantum annealing solution for this problem, formulated within a constrained quadratic model (CQM) framework. This approach delivers solutions significantly faster than conventional methods, offering a performance advantage that can translate into measurable cost savings and improved operational efficiency.
We bring a unified and modular AI/Quantum intelligence to the maritime logistics sector—blending macroeconomic modeling, autonomous navigation, and vessel-level decision intelligence. Whether optimizing global trade foresight, enabling unmanned vessel control, or orchestrating container flows in port, our systems are built for operational robustness, regulatory alignment, and sustainable growth.
We enable maritime stakeholders to shift from reactive logistics to proactive, intelligence-driven control—ensuring that shipping networks remain adaptive, efficient, and future-ready in the face of economic uncertainty and environmental constraints.