Research Article

Partition Learning for Multiagent Planning

Figure 21

Sample sequence of partition learning classification path planning for a scenario involving six agents and six targets (the number of targets was unknown to the agents). Agents are represented by blue/green circles with protruding lines that represent the direction of the agents. The approximate field of view of each agent is represented by green connected lines. Targets are represented by black/red circles. The entire surveillance area is partitioned. These partitions are represented by different colors. Follow the sequence from left to right and from top to bottom. In (a) initial partitions are formed. In (b) tracking partitions appear. By (f) all targets are within tracking partitions.
590479.fig.0021a
(a)
590479.fig.0021b
(b)
590479.fig.0021c
(c)
590479.fig.0021d
(d)
590479.fig.0021e
(e)
590479.fig.0021f
(f)