Fixed wing is a mission hypothesis to test, not the default answer
Wide-area reconnaissance often favors continuous cruise: no braking or hovering, with energy spent mainly on forward motion and coverage. Fixed-wing aircraft also impose minimum airspeed, turn radius, camera field-of-view and handoff-geometry constraints. Multirotor or hybrid aircraft may still be better for confined terrain and repeated close observation. The simulator should compare these options instead of declaring a winner in advance.
A target is not a permanently precise point on the map
A scout actually provides a timestamped measurement: position, estimated velocity, line-of-sight geometry, observation quality and classification confidence. After the aircraft departs, the system may infer where the target could be from that evidence; it must not continue reading hidden simulator truth.
The current ground-contact research overlay branches candidate headings from nearby public road geometry. It is not yet a connected road graph with intersection-transition probabilities. The maritime panel shares the Python reference model's last heading, speed, synthetic current and 2.5% wind-drift assumptions, forming three expanding elliptical branches. With no new evidence for 45 seconds it states LOST. The display keeps “last measured” separate from “estimated now.”
Negative evidence still updates probability after visibility is lost
Buildings, terrain, sea state, viewing angle or communications can hide a target temporarily. The track then enters COASTING: confidence falls and the distribution widens over time. If another scout clearly searches a predicted region without detecting the contact, that branch is down-weighted rather than arbitrarily set to zero.
- PROPOSEDPropose a handoff with geometric evidence
- ACCEPTED / EXECUTINGThe incoming scout passes capability checks and begins
- FOV OVERLAP ACKTwo real sensor sectors overlap
- COMPLETE / LOSTRelease the predecessor—or state loss honestly
The current handoff model uses both aircraft positions, headings, sensor half-angles and ranges to verify common field of view. The fixed-wing candidate also pays turn time derived from cruise speed and bank angle, and passes an altitude-aware building-path check. Only an incoming scout in EXECUTING with a common-FOV acknowledgement allows the state machine to release the predecessor.
The best path may curve, but it should not weave merely for appearance
A fixed-wing coverage route will often contain long scan legs, shallow arcs and turn-radius-constrained reversals. The research build already brings bank-angle-limited tracking turns and obstacle-height checks into handoff feasibility. Public aircraft positions still come from an immutable replay; full receding-horizon cost optimization remains a next step. Every additional turn must eventually justify itself through information gain, coverage or collision-avoidance safety.
The useful comparison is not simply straight versus curved. It is the combined tradeoff among new area covered, reacquisition probability, handoff gaps, turning energy and safety constraints.
Questions the simulator should answer
- Does fly-through handoff preserve more useful coverage than stopping to track?
- Does the incoming scout obtain enough overlap before the outgoing scout leaves?
- Is the probability distribution calibrated instead of overconfident?
- How quickly can the system reacquire after sensor, link or aircraft loss?
- How much new coverage and uncertainty reduction are obtained per unit of modeled energy?
A fixed seed enables a fair replay; it does not make the real world perfectly predictable.