3D Path-finding in a voxelized model of an indoor environment

More Info
expand_more

Abstract

Indoor environments can be complex and need to be represented by a 3D representation. This is especially important for 3D path-finding. 3D features of the environment can have an effect on the computed path and these features can not be represented correctly in a 2D representation. A voxelized model is a good 3D representation that supports path-finding. It is a geometrical and topological model and it is easy to incorporate semantics. It is also fairly easy to consider the shape and size of the actor. This thesis presents a new path-finding method that operates on a voxelized model and supports different kinds of actors. Distinction is made between actors by their size and mode of locomotion. Supported modes of locomotion are walking, driving and flying. A hierarchical data structure is used to reduce the time complexity of the path-finding problem. This hierarchical data structure is a graph which is derived from a cell decomposition of space. The results indicate that the path-finding method in its current state operates well for walking and driving actors, but further improvements are required for flying actors.