Search papers, labs, and topics across Lattice.
This paper introduces PBD-R, a revised Position-Based Dynamics (PBD) formulation that enforces physically accurate rigid-body dynamics by incorporating a novel momentum-conservation constraint and a modified velocity update. They also present a solver-agnostic benchmark with analytical solutions for evaluating physical accuracy in rigid-body simulations. Results demonstrate that PBD-R significantly outperforms standard PBD and achieves accuracy comparable to MuJoCo, but with lower computational cost.
Achieve MuJoCo-level accuracy in rigid-body dynamics using a particle-based method (PBD-R), opening the door to unified simulations of rigid, deformable, and fluid materials in robotics.
Robotics demands simulation that can reason about the diversity of real-world physical interactions, from rigid to deformable objects and fluids. Current simulators address this by stitching together multiple subsolvers for different material types, resulting in a compositional architecture that complicates physical reasoning. Particle-based simulators offer a compelling alternative, representing all materials through a single unified formulation that enables seamless cross-material interactions. Among particle-based simulators, position-based dynamics (PBD) is a popular solver known for its computational efficiency and visual plausibility. However, its lack of physical accuracy has limited its adoption in robotics. To leverage the benefits of particle-based solvers while meeting the physical fidelity demands of robotics, we introduce PBD-R, a revised PBD formulation that enforces physically accurate rigid-body dynamics through a novel momentum-conservation constraint and a modified velocity update. Additionally, we introduce a solver-agnostic benchmark with analytical solutions to evaluate physical accuracy. Using this benchmark, we show that PBD-R significantly outperforms PBD and achieves competitive accuracy with MuJoCo while requiring less computation.
