Researchers at the Massachusetts Institute of Technology (MIT) have trained an AI model to optimize movements within a warehouse environment.
When an order is received, a robot is dispatched to a designated area, retrieves the required item from a shelf, and delivers it to a human operator. Hundreds of mechanical assistants perform this task simultaneously, and if their paths intersect, they risk damage.
Traditional search-based algorithms prevent potential collisions by holding one android in place while altering the trajectory of another. However, as their numbers increase, the optimization challenge grows exponentially.
Scientists observed that moving robots resemble cars trying to find the best route in a crowded city center.
They developed a deep learning model that encodes crucial information about the warehouse, including mechanical loaders, planned routes, tasks, and obstacles. The neural network uses this data to identify suitable areas of the warehouse that should be cleared.
“We have developed a new architecture that encodes hundreds of robots in relation to their trajectories, destinations, and interactions with each other,” said Cathy Wu, assistant professor of civil and environmental engineering at MIT.
Beyond optimizing warehouse movements, this deep learning method can be applied to other complex planning tasks, such as designing computer chips or laying pipes in large buildings.
Earlier, ForkLog in its News+ format reported on androids being developed for work in factories and warehouses.