Autonomous Cave Rescue: Frontier Exploration & Multi-Robot Path Planning (2D)
2D autonomous cave-rescue lab: a real frontier-exploration algorithm drives a fleet of robots through a procedurally generated cave, building an occupancy grid from simulated LIDAR raycasts and planning BFS paths to the nearest unclaimed frontier until every trapped survivor is found.
This 2D companion replaces the 3D version's decorative scenery with the actual algorithm a real rescue-robot fleet runs: a fog-of-war occupancy grid built from simulated LIDAR raycasts, breadth-first frontier search that always sends each robot to the nearest unclaimed unexplored boundary, and survivor detection gated on genuine line-of-sight rather than proximity alone. Cave density, fleet size, sensor range and robot speed are all live parameters that change how the search behaves, not cosmetic sliders.
Cellular-automaton cave generation with guaranteed connectivity, a shared occupancy grid from simulated LIDAR raycasting, nearest-unclaimed-frontier BFS path planning per robot, and line-of-sight-gated survivor discovery.
2D · HTML5 Canvas 2D · 60 FPS target · runs fully client-side, no install