MIT Presents Drone Path That Guarantees Collision Avoidance in Unknown Environments

According to AI TIMES, MIT researchers have developed an autonomous path-planning technology called SANDO that mathematically guarantees a drone flight path that avoids moving obstacles and does not collide, even in an unmapped unknown environment. In 12 real drone tests, the system relied only on the onboard computer to recalculate a safe path in response to changes in the surroundings, avoiding all dynamic obstacles, and the research findings were disclosed on the 26th of last month. The core of the technology is not to try to predict the exact future movements of obstacles, but to continuously update safe corridors by calculating which spaces will still be empty in the future based on their maximum speed. It also adds a method for finding a more efficient route while avoiding obstacle-congested areas, aiming to improve both safety and mobility performance. The researchers said it could help drones continue search and mission operations in places where maps are hard to trust, such as wildfires, collapsed buildings, and mines. It also leaves room for expansion into autonomous flight technology that can explain the validity of its route in places where predictions often go wrong, such as delivery and complex urban environments.