Teaching robots to map large environments
MIT researchers have developed a new AI-driven system that allows robots to generate 3D maps of large, complex environments in seconds by stitching together smaller submaps using classical geometry.
A robot searching for workers trapped in a partially collapsed mine shaft must rapidly generate a map of the scene and identify its location within that scene as it navigates the treacherous terrain. Researchers at MIT have developed a new system that draws on ideas from both recent artificial intelligence vision models and classical computer vision to process an arbitrary number of images. Their system accurately generates 3D maps of complicated scenes in a matter of seconds. The AI-driven system incrementally creates and aligns smaller submaps of the scene, which it stitches together to reconstruct a full 3D map while estimating the robot’s position in real-time. Unlike many other approaches, their technique does not require calibrated cameras or expert tuning. Beyond search-and-rescue, this method could be used for extended reality applications or industrial robots in warehouses. The researchers demonstrated that it is possible to generate an accurate 3D reconstruction with an average error of less than 5 centimeters using only short videos captured on a cell phone.
Source: Robohub
