I want to implement a physics simulator for large floating base rigid body systems from scratch. The Rigid Body Dynamics Systems (RBD) should typically have the following characteristics:

  1. About ~50 degrees of freedom (DoF) plus the 6 DoF of the floating base (e.g., a legged robot, a humanoid robot, etc.).
  2. It will always have tree structure (no parallel robots)
  3. The simulation needs to be fast.
  4. There will be contact forces (e.g., during walking, grasping or manipulation)

What would be the best physics formulation in terms of numerical accuracy and computational complexity? Lagrange appears too slow, but Newton Euler is often hard to comprehend.

  • $\begingroup$ What have you tried? What have you read? The problem you want to solve is in the "fluid-structure interaction" category, for which there is a huge literature. Of course, none of the methods describe are both accurate and fast. $\endgroup$ Commented Aug 12, 2019 at 12:41
  • $\begingroup$ I have tried Newton-Euler and Lagrange for Robot Arms based on Yoshikawa's 1989 book. There must be better things now! $\endgroup$ Commented Aug 14, 2019 at 14:26


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.