MIT researchers have developed a new method of training drones to navigate by using a “liquid neural network” that can simulate human learning. The system is capable of learning from a small amount of data and can adapt to new environments quickly.

According to MIT postdoctoral associate William Gilpin, “The system is inspired by the way the human brain works. It’s designed to mimic the brain’s ability to learn from experience and adjust to new situations.”

The liquid neural network uses a “memristor-based neural network” that is trained to recognize patterns in visual data. The network is then used to control the drone’s flight path, allowing it to navigate through complex environments with ease.

MIT Professor Joel Voldman explained that “the memristor-based neural network allows the system to learn from a very small amount of data, which is important for drone applications where collecting large amounts of data can be difficult.”

The researchers believe that their system could be used for a variety of applications, including search and rescue operations and environmental monitoring. “The system has the potential to revolutionize the way we use drones for a variety of applications,” said Voldman.

This new method of training drones could be a game-changer for the industry, as it allows drones to learn and adapt to new environments quickly and efficiently. As Gilpin put it, “Our system enables drones to learn like humans do, allowing them to navigate through complex environments with ease.”

By The Impactlab