MIT researchers reported a flight-planning method called SANDO that aims to keep a drone clear of moving obstacles in an unfamiliar space. The team said its onboard system avoided dynamic obstacles in 12 physical test flights. Source: MIT News.
Planning around motion
SANDO builds a time-sensitive safe corridor rather than assuming obstacles stay where sensors first detected them. It estimates how far an obstacle could move based on a speed bound and replans the aircraft’s path as the scene changes. That addresses a practical gap in autonomous flight through cluttered spaces.
What was demonstrated
MIT reported collision-free results in simulations and 12 real UAV flights using onboard computing and sensors. The mathematical safety claim applies under the model and assumptions used by the researchers; it should not be read as a guarantee for every sensor failure or real-world environment.
Why this could matter
Disaster response, indoor inspection and delivery routes all present changing obstacles. The next evidence to watch is performance in larger, less controlled spaces, computational load on small aircraft and how the planner behaves when detection or speed estimates are wrong.
Share:
Atlanta Fire Drones Pass 880 Calls
Beam Signs ScoutDI Acquisition Deal