Parallel multibody solver coupling algorithms

This project concerns the efficient simulation of constrained-multi body systems with applications in training simulations. For instance, a crane on a construction site can be modeled and simulated as a collection of rigid bodies connected by rotational joints. Simulation of contact and friction is similar but a challenge because the force is bounded (i.e., forces are not allowed to act like glue and can only push objects apart). When there are large numbers of bodies in a simulation, with many frictional contacts, these systems can be challenging to
solve. In this work, we allow the systems to be partitioned, solved in parallel, and then coupled with a solve of the interface forces. This work specifically aims to improve the efficiency of this interface solve, with the main benefit being that much larger training systems can be solved both accurately, and at speeds that are useful for interactive training simulations.

Faculty Supervisor:

Paul Kry;Sheldon Andrews

Student:

Henry Ho

Partner:

CMLabs

Discipline:

Computer science

Sector:

Information and cultural industries

University:

Program: