Particle-Based Simulation of Rigid Bodies for Real-Time Applications

  • Solanillas Francés, David Manuel (McGill University)
  • Kövecses, József (McGill University)

Please login to view abstract download link

Particle-based methods are well suited for contact-rich problems such as granular flows, debris motion, and demolition, where large numbers of interacting bodies must be handled efficiently. Their scalability and robustness make them attractive alternatives to classical rigid-body approaches, which often struggle in dense contact scenarios. This work is based on the Parallel Particles (P²) Method~\cite{DH} and extends it by introducing constant distance constraints between particles. These constraints are used to form groups of particles that behave as rigid bodies. Contact interactions are modelled exclusively through sphere–sphere overlap, resulting in a simple and robust contact formulation. The resulting approach preserves the fully parallel nature of the P² method, as both contact and internal constraints are solved within a unified Jacobi-style framework using mass splitting and averaging. No additional rigid-body solvers or complex contact geometries are required. The method is demonstrated through real-time simulations of stable structures, which can be subsequently demolished under external loading (Fig.~\ref{fig:rigid_aggregates}(a)). After collapse, large amounts of debris are handled efficiently despite high contact densities, while maintaining numerical stability and interactive performance. As a future perspective, the same bilateral constraint formulation can be used to define different composite particle shapes (Fig.~\ref{fig:rigid_aggregates}(b)), enabling the construction of tailored particle morphologies with potential applications in soil–structure interaction problems such as pile–soil simulations.