Efficient Yarn-Based Cloth with Adaptive Contact Linearization

How to read this

Category

Method: an acceleration technique for yarn-level cloth simulation

Contributions

• Adaptive contact linearization that computes exact contact response at one step then approximates forces in nearby configurations with a rotated linear force model
• Exploitation of temporal coherence of internal contacts so the approximation is updated only infrequently, adaptively in space and time
• Contact-model tracking that reduces detection time, yielding a 7-9x speedup in contact processing and 4-5x overall, enabling character-scale garments

Context

Directly extends yarn-level cloth simulation (Kaldor et al., Simulating Knitted Cloth at the Yarn Level) by attacking the contact-processing bottleneck in densely interlinked geometry.Builds on: Simulating Knitted Cloth at the Yarn Level

Correctness

The key assumption is that internal cloth contacts deform with enough temporal coherence that a linearized force model stays accurate between infrequent updates; readers should note the approximation is validated on knitted character garments and accuracy depends on how adaptively updates are scheduled.

Clarity

Moderately technical; a first pass conveys the linearize-and-reuse idea, do a second pass for the rotated linear force model and the adaptive update criteria.

How to read it

Focus on why yarn-yarn contact is the bottleneck and on the rotated linear force approximation; a second pass on the adaptive update scheduling pays off if you care about the accuracy/speed trade-off.