SYNTHETIC ILLUSTRATIVE SIMULATION Procedural world, simulated drone and subject. Not a real drone or flight controller, not validated 3D performance, and unrelated to BeliefView v2 metrics (the frozen v2 experiment is 2D).

BeliefView 3D When the sighting ends, the search begins.

Spectator pane: evaluation-only spectator camera (not available to the drone), optional cutaway  ·  Drone pane: simulated drone camera; cover stays opaque
TRACKING
Active guidance: Kalman
—
Observing

Flight & sensor diagnostics
Drone mode
—
Altitude (AGL)
—
Ground speed
—
Since last fix
—
Prediction radius (95%)
—

Sensor model (simulation)

Line of sight
—
In camera FOV
—
Range
—
Explore scenario & controls

Deterministic replay

All policies receive the exact static synthetic map in advance. Cover and terrain are simplified opaque geometry, not discovered or reconstructed by the drone.

Display overlays

Overlays

Search guidance

Both policies start from the same coarse area. The subject reacts to each drone path, so changing guidance can change the subject trajectory. This is not a fixed-path counterfactual.

Model evidence & probability overlay
Model: not loaded (Kalman guidance selected)

Learned search scores candidate low viewpoints by map-visible probability mass, minus travel cost, with a preference for the current viewpoint. This is a greedy heuristic, not an exact probability of reacquisition. A plan can fall back to the Kalman viewpoint.

Synthetic: the learned model is trained only on simulated episodes of a hand-written evasive subject. It is an illustrative demo, not validated on real people, drones or forests, and its probabilities are not real-world likelihoods.

Belief map · tracker view
0
10
20
30 m
canopy layer
● DRONE CAM