MIT researchers have developed an AI-powered system to design autonomous underwater gliders inspired by marine animals. These gliders are more efficient and have novel shapes that outperform traditional torpedo-shaped designs. The researchers used machine learning to test different 3D designs in a physics simulator, then molded them into more hydrodynamic shapes.
They created a dataset of conventional and deformed shapes, simulating how they would perform at different angles-of-attack. A neural network evaluated the lift-to-drag ratio of each shape, identifying those most likely to glide efficiently through water. The AI-designed gliders have better lift-to-drag ratios, exerting less energy and moving more efficiently across a pool.
The gliders come in unique shapes, such as a two-winged machine resembling an airplane and a four-winged object resembling a flat fish with four fins. These shapes are a departure from the traditional torpedo-shaped designs, and their efficiency is a testament to the power of AI in design optimization.
The gliders can be fabricated via 3D printing, using significantly less energy than handmade ones. This technology has the potential to revolutionize ocean exploration by enabling more efficient data collection on currents, salt levels, and climate impacts. With their improved efficiency and maneuverability, these AI-designed underwater gliders could help scientists gather more detailed insights about the ocean and monitor the impacts of climate change
