Exact Cell Decomposition

Abstract

In this chapter we describe a second approach to motion planning, exact cell decomposition. The principle of this approach is to first decompose the robot’s free space Cfreeinto a collection of non-overlapping regions, called cells, whose union is exactly1 Cfree(or its closure). Next, the connectivity graph which represents the adjacency relation among the cells is constructed and searched. If successful, the outcome of the search is a sequence of cells, called a channel, connecting the cell containing the initial configuration to the cell containing the goal configuration. A path is finally extracted from this sequence.