On approximate geodesic-distance queries amid deforming point clouds

Abstract

We propose data structures for answering a geodesic-distance query between two query points in a two-dimensional or three-dimensional dynamic environment, in which obstacles are deforming continuously. Each obstacle in the environment is modeled as the convex hull of a continuously deforming point cloud. The key to our approach is to avoid maintaining the convex hull of each point cloud explicitly but still able to retain sufficient geometric information to estimate geodesic distances in the free space. © 2009 Springer-Verlag.