Guided Exploration with Dynamic Potential Fields: the Cubical Path System

Abstract

Exploring unknown models or scenes is a highly interactive and dynamic process. Systems for automatic presentation of models or scenes either require cinematographic rules, direct human interaction, framesets, or pre‐calculation of paths to a known goal. In this paper we present a system which can deal with rapidly changing user interest in objects of a scene or model as well as with dynamic models and changes of the camera position introduced interactively by the user or through cuts. We describe CubicalPath, a new potential field‐based camera control system that helps with the exploration of virtual environments.

Camera navigation support in a virtual environment

“…Local potential fields coming out of obstacles form unpredictable extremes which are situated in accidental places and they are blocking the camera on the way to the aim [45,49]. Forces coming out of potential field distribution do not let the camera approach obstacles.…”

Tour generation for exploration of 3D virtual environments

“…CubicalPath [Beckhaus et al 2001] uses a similar off-line voxelization of the 3D world and then employs dynamic potential fields for guiding exploration at run-time.…”

Designing quality walkthroughs

“…One of the motivations behind this choice is to effectively utilize some of the data structures that we have built for the tour planning problem. The reader is encouraged to refer References [10,16] for a detailed discussion on APF for navigation. In this section we discuss how it is adapted to our context.…”

“…[3][4][5][6][7][8][9] We use an artificial potential field (APF)-based method to dynamically compute a local path which merges with the tour. Our approach is similar to that of Beckhaus et al 10 However, we take advantage of a pre-computed visibility data structure. This allows us to avoid the costly pre-computation and storage of the potential field values at discretized voxels, which are determined efficiently at runtime.…”

“…This allows us to avoid the costly pre-computation and storage of the potential field values at discretized voxels, which are determined efficiently at runtime. In Reference [10] one needs to identify explicit navigation way-points to avoid problems of local minima. In our approach the notion of way-points is different.…”