Johns Hopkins Uses AI to Model Safer Drone Traffic Systems
-
Johns Hopkins researchers used AI to model safer drone traffic systems to handle expected growth in drone flights.
-
The model uses autonomous algorithms for collision avoidance and strategic deconfliction to enhance safety.
-
Realistic "noisy sensor" and risk level algorithms make the system adaptable like real-world conditions.
-
Simulations tested scenarios including rogue drones deviating from planned routes with promising safety results.
-
This research builds on over 20 years of work by Johns Hopkins on enhancing US National Airspace System safety.