Enhanced Waypoint Graph for Path Planning in Virtual Worlds

Our research focuses on the problem of path planning, which often occurs in virtual world applications. We propose an automatic generation of enhanced waypoint graph, which is a graph data structure consisting of point nodes, which describe the corner features in the virtual world, as well as edges connecting those nodes. Given a polygon soup representation of a virtual world, for every character radius, the proposed algorithm starts by constructing a discrete distance field, consisting of regularly sampled points in 3D space. Corner detection and clustering are then done with respect to the points whose distance values are slightly larger than the character size to get the waypoints. These waypoints are further sparsely connected using traversability test, taking into account their distances to nearby obstacles. The resulting enhanced waypoint graph is sparse but has regularly distributed edges emanating from each waypoint. In addition, the graph is also able to handle different types of motions for characters with various sizes.