Different worlds.
Testable questions.

Do clear peer messages help, and what happens when they are scrambled?

How do policies respond to forged reports, unreliable modalities and uncertainty? The rescue setting is a simulated test world, not deployment validation.

Can local distance, light or environmental readings change a node’s contribution to the group? Optional sensors need electrical and firmware integration.

Can speakers and onboard microphones let neighbouring nodes exchange cues? Timing, interference and feedback need measurement.

What patterns appear when local rules influence sound, light or rhythm across the group? A compelling performance is a starting point for a test.