This video shows human machine optimal decision making for sequential inspection task where static ground targets as classified as either being threats or non-threats. The operator is tasked with identification of static targets in the video imagery live-streamed from four UAVs. We demonstrate a significant reduction in the false alarm rate of the closed loop system over a "human only" open loop classification system.
This video shows two UAVs autonomously patrolling a perimeter and inspecting alert sites/stations. The imagery collected at the alert stations is sent to a remote (ground) operator for classification. Each UAV decides which station to visit next and also how much time to spend inspecting an alert site. The flight paths are based on true telemetry data from flight tests.
This video shows results from a successful flight test (demo) of 2 UAVs autonomously making decisions to isolate (capture on camera) a moving ground target with limited ground based sensing of the target motion. The UAVs share information and collaborate to quickly chase down and capture the intruder.