Neural dynamics-based Poisson propagation for deformable modelling

Jinao Zhang, Yongmin Zhong, Julian Smith, Chengfan Gu

Research output: Contribution to journalArticleResearchpeer-review

4 Citations (Scopus)

Abstract

This paper presents a new methodology from the standpoint of energy propagation for real-time and nonlinear modelling of deformable objects. It formulates the deformation process of a soft object as a process of energy propagation, in which the mechanical load applied to the object to cause deformation is viewed as the equivalent potential energy based on the law of conservation of energy and is further propagated among masses of the object based on the nonlinear Poisson propagation. Poisson propagation of mechanical load in conjunction with non-rigid mechanics of motion is developed to govern the dynamics of soft object deformation. Further, these two governing processes are modelled with cellular neural networks to achieve real-time computational performance. A prototype simulation system with a haptic device is implemented for real-time simulation of deformable objects with haptic feedback. Simulations, experiments as well as comparisons demonstrate that the proposed methodology exhibits nonlinear force–displacement relationship, capable of modelling large-range deformation. It can also accommodate homogeneous, anisotropic and heterogeneous materials by simply changing the constitutive coefficient value of mass points.

Original languageEnglish
Number of pages11
JournalNeural Computing and Applications
Volume31
Issue numberSuppl 2
DOIs
Publication statusPublished - Feb 2019

Keywords

  • Cellular neural networks
  • Deformable objects
  • Nonlinear deformation
  • Poisson equation
  • Real-time performance

Cite this