MIT drone planner avoids moving obstacles in 12 test flights

MIT's SANDO navigation system avoided all moving obstacles across 12 real drone flights, New Atlas reports. In simulations, MIT says it also reached destinations faster than several comparison systems while avoiding collisions in every environment tested. The system navigates without a prior map by tracking obstacles and calculating every position they could reach within a maximum speed. It builds a corridor through free space and continually updates the drone's route using its onboard computer. Its mathematical collision-avoidance guarantee depends on obstacles staying within that speed limit. The flight tests do not establish readiness for every disaster scenario.