Multi-Resolution Hierarchical Motion Planner for Multi-Robot Systems on Spatiotemporal Cellular Automata

This chapter presents a new multi-resolution and hierarchical approach to the problem of motion planning of Multi-Robot Systems on discretized spaces. The goal is to operate on large spaces (compared to the size of the robots), where the number of cells quickly becomes untreatable, in particular for interacting robots problems, without losing precision (resolution). To work around this problem, we have introduced 3 levels of maps: the first is topological, the second a rectangular tessellation covering the free space, and the third a regular (small) cells decomposition. The first two maps are used to reduce the problem and to simplify it with non-accurate planning. Limiting the search space to smaller areas of nterest at the last level and considering the interactions between robots, precise parallel motion planning is performed using Spatiotemporal Cellular Automata