As subsea robotics become more capable and autonomous, the way we develop, validate, and train to use them must advance too. High-fidelity, physics-based simulation moves beyond traditional training, creating true-to-nature virtual environments where engineering, rapid prototyping, personnel development, and AI developments can progress together.
For engineers, simulation provides a powerful rapid prototyping and proving environment. ROV configurations, tooling, subsea infrastructure, environmental conditions, and operational constraints can be introduced virtually, enabling concepts to be built, tested, challenged, and refined using simulation before physical prototypes or offshore trials. The presentation will show an example project where a suite of tools was designed and tested in simulation prior to tank SIT and eventual use offshore, including the revision of tooling designs and project plans.
The simulation environment enables ROV pilots and technicians to train against realistic representations of the systems, tasks, and conditions they may encounter offshore. Personnel can rehearse complex operations, familiarize themselves with new tooling, and refine techniques alongside technology development. An example training scenario demonstrating the impact of changing environmental conditions will be presented. The same environment used to train personnel is being used to train and validate AI supervised behaviors. This presentation will also a sample case where simulation was used to evaluate an autonomous vehicle’s behavior subsea for troubleshooting after a real-world campaign.
Simulation also addresses a major challenge for AI and machine learning: access to high-quality subsea training and validation data. This presentation will present a case where an object identification, pose recognition, and supervised autonomous behavior for a work class ROV were developed using synthetic data generated in simulation while the real-world data lacked validation that can only be provided in simulation.
This presentation explores how simulation-driven engineering connects rapid prototyping, human performance, AI development, and synthetic data created in a virtual environment improve subsea outcomes for ROVs and AUVs.