6-DoF Haptic Rendering Using Continuous Collision Detection between Points and Signed Distance Fields.

Area of Science:

• Computer Graphics
• Robotics
• Human-Computer Interaction

Background:

• Continuous collision detection is crucial for realistic simulations in animation, virtual reality, and haptics.
• Existing methods are primarily limited to polygonal geometry, hindering applications with complex shapes.
• Signed distance fields offer a powerful representation for complex geometry.

Purpose of the Study:

• To develop a fast and robust continuous collision detection algorithm for point-based representations and signed distance fields.
• To enable 6-DoF (six degrees of freedom) haptic rendering of contact for objects with complex geometry.
• To compare different haptic rendering methods for point-versus-distance field contact.

Main Methods:

• An octree subdivision of the signed distance field for efficient cell traversal.
• A method for resolving continuous collisions between point clouds and signed distance fields.
• Investigation and comparison of continuous integration of penalty forces and constraint-based methods for haptic rendering.

Main Results:

• Robust and continuous detection of intersections between points and signed distance field level sets.
• Successful rendering of contact between rigid objects with complex geometry using point clouds and signed distance fields.
• Experimental validation showing the continuous collision detection method outperforms discrete methods, especially at high velocities.

Conclusions:

• The proposed algorithm provides a robust solution for continuous collision detection with signed distance fields.
• This enables advanced 6-DoF haptic rendering for complex geometries, previously unachievable.
• The continuous approach enhances simulation stability and realism in interactive applications.