Add a terrain-walker pathfinding algorithm that generates human-like paths through terrain, rather than a globally optimal shortest path.
Starting from A, the walker:
- Looks toward B and analyzes the locally visible terrain.
- Selects a plausible intermediate waypoint, favoring natural features such as valleys, saddles, passes and gentle slopes.
- Uses Dijkstra to reach that waypoint.
- Reassesses the terrain from the new position and selects the next waypoint.
- Repeats until reaching B.
Unlike standard Dijkstra/A*, the walker should not have global knowledge of the terrain. The path emerges from a sequence of local decisions, potentially producing a longer but more plausible trail.
This could be particularly useful for procedural road, trail and settlement generation.
Add a
terrain-walkerpathfinding algorithm that generates human-like paths through terrain, rather than a globally optimal shortest path.Starting from A, the walker:
Unlike standard Dijkstra/A*, the walker should not have global knowledge of the terrain. The path emerges from a sequence of local decisions, potentially producing a longer but more plausible trail.
This could be particularly useful for procedural road, trail and settlement generation.