A Bi-Level Probabilistic Path Planning Algorithm for Multiple Robots with Motion Uncertainty

Figure 6

Generation of policy for in each district. Assuming that the candidate actions in each district include , , and , the generation of policy for robot is shown in this figure. In this tree structure, different level of branches corresponds to different steps. Each circle represents the action for in the district at the corresponding time step. For example, the action left means that all the robots in the district should go left. A complete policy from the first time step to the αth time step is illustrated in the blue box. Multiple policies for each robot can be generated from this tree structure.