A Framework for Real-Time Nonholonomic Path Planning in Huge Terrain Datasets

H. Thwaites (Editors)
In proceedings of The 11th International Conference on Virtual Systems and Multimedia (VSMM2005), pages 583-592, ARCHAEOLINGUA, Oct. 2005
Presented at The 11th International Conference on Virtual Systems and Multimedia (VSMM 2005)
 

Abstract

When users navigate through a terrain dataset, they typically control the camera directly. However, such control is difficult and often results in rather unsteady camera motions that can easily cause motion sickness, especially in virtual reality. In this paper we describe a new technique for automatic generation of a camera motion path using nonholonomic motion planning techniques. Using our approach, the user simply specifies an object of interest, e.g. on a map or from a list of landmarks, together with the desired viewing distance associated with that object. Then the system automatically computes a smooth camera motion from the current position and orientation to show the specified point. Our framework allows the simple implementation of different navigator metaphors that specify the camera behaviour. Due to the size of recent terrain datasets and the complexity of nonholonomic motion planning, using standard approaches is not possible. To solve this problem, we present a multi-scale approach using a simple and fast path planning to find coarse resolution paths and iteratively refine them. Although developed for terrain exploration, our framework is general enough to be used in any virtual environment. One of the key advantages of our method is, that the user can take over control at any time, e.g. to examine a close-by object, and switch back to automatic navigation later. This way our path planning framework can also be regarded as navigation assistant.

Download Paper

Download Paper

Bibtex

@INPROCEEDINGS{guthe-2005-framework,
      author = {Guthe, Michael and Klein, Reinhard},
      editor = {Thwaites, H.},
       pages = {583--592},
       title = {A Framework for Real-Time Nonholonomic Path Planning in Huge Terrain Datasets},
   booktitle = {The 11th International Conference on Virtual Systems and Multimedia (VSMM2005)},
        year = {2005},
       month = oct,
   publisher = {ARCHAEOLINGUA},
    abstract = {When users navigate through a terrain dataset, they typically control the camera directly. However,
                such control is difficult and often results in rather unsteady camera motions that can easily cause
                motion sickness, especially in virtual reality. In this paper we describe a new technique for
                automatic generation of a camera motion path using nonholonomic motion planning techniques. Using
                our approach, the user simply specifies an object of interest, e.g. on a map or from a list of
                landmarks, together with the desired viewing distance associated with that object. Then the system
                automatically computes a smooth camera motion from the current position and orientation to show the
                specified point. Our framework allows the simple implementation of different navigator metaphors
                that specify the camera behaviour. Due to the size of recent terrain datasets and the complexity of
                nonholonomic motion planning, using standard approaches is not possible. To solve this problem, we
                present a multi-scale approach using a simple and fast path planning to find coarse resolution paths
                and iteratively refine them. Although developed for terrain exploration, our framework is general
                enough to be used in any virtual environment. One of the key advantages of our method is, that the
                user can take over control at any time, e.g. to examine a close-by object, and switch back to
                automatic navigation later. This way our path planning framework can also be regarded as navigation
                assistant.},
        isbn = {963 8046 63 5},
  conference = {The 11th International Conference on Virtual Systems and Multimedia (VSMM 2005)}
}