Case Study: University of Southern Mississippi advances UUV training with COMSOL simulation apps

University of Southern Mississippi builds UUV simulation apps with Comsol in 1-2 hours

The University of Southern Mississippi (USM) faced challenges in training uncrewed underwater vehicles (UUVs) to accurately detect ferromagnetic objects on the ocean floor due to platform noise, environmental clutter, and the high cost of physical scans. To overcome this, they partnered with vendor Comsol and used the COMSOL Multiphysics® simulation platform along with its AC/DC Module and Uncertainty Quantification Module to build and validate potential field models.

Comsol's solution enabled USM to create simulation apps that generated large, high-quality synthetic datasets to train machine learning algorithms. This approach provided a rich training set with over 500,000 data points, allowing the data science team to develop models for automatic target recognition. The result was a blueprint for integrating simulation-driven machine learning into UUV operations, significantly improving predictive capabilities and paving the way for more intelligent and accurate seafloor mapping.

University of Southern Mississippi