A Unified Framework for Posing Strongly-Coupled Multiphysics for Robotics Simulation

Jeong Hun Lee, Junzhe Hu, Sofia Kwok, Carmel Majidi, Zachary Manchester

Keywords: Multiphysics, Non-Rigid Simulation, Robotics

TL;DR: Unified framework for deriving and simulating multiphysics applied to non-rigid fluid-robot interactions

Abstract: We present a framework for posing multiphysics as a single, unified optimization problem for downstream robotics simulation. Specifically, we apply our framework to the fluid-robot interaction setting, where the governing coupled manipulator and incompressible Navier-Stokes equations are derived together from a single Lagrangian using the principal of least action. We then employ discrete variational mechanics to derive a stable, implicit time-integration scheme for jointly simulating both the fluid and robot dynamics, which are tightly coupled by a constraint that enforces the no-slip boundary condition at the fluid-robot interface. We demonstrate our approach's physical accuracy on a novel swimming robot in simulation and validate results on real-world hardware, showcasing our framework's sim-to-real capability. In future work, we hope to extend our unified framework to pose other multiphysics settings commonly found in robotics, such as the manipulation of deformable objects.

Submission Number: 2