Figure 2.

Example grid divided up into 3 connected regions or nodes. The agent’s motion model consists of different regions (also referred to as nodes). Each region contains a subset of cells (white, grey or checkered). A cell may belong to multiple regions, not shown. Each region has a central point (black circle) contained within the cell. The actions available to the agent at a given region is to either search the current region, or move to an adjacent region along a path connecting central points.