Hi, thanks for the suggestions! The thing is that I use convex cast to detect threatenings for obstacle avoidance and I don’t really want things smashed and physically resolve collision. I use A* for pathfinding and try to use nape to do obstacle avoidance for moving entities.
I don’t quite get the optimization as not doing obstacle detection on every frame though...as entities are moving on stage, how am I going to determine the obstacle detection step? It seems not realistic as all entities speed and route all needs to be counted...