I'm trying to compute the inverse dynamics of a biped (hybrid dynamics) and I have a lot of questions. Could someone help me with some information?
Well, I know that I can use recursive Newton Euler or Lagrangians Dynamics. But since I don't know how to compute the matrix that contains the Coriolis and Centripetal terms, I'm using RNEA. What are the best papers that explain Lagrangina dynamics and how to compute this matrix?
Another question is how to compute the inverse dynamics problem when the biped is in the double stance phase. Since in this phase I have to solve a closed loop (I'm not sure if this is always true, since I think that it's depend of the contact model used in the simulation). Lagrangian method solve this problem?
For the forward simulation, I don't know what is the best approach. I prefer use PGS method like Bullet or LCP method like ODE, but I've found implementations where penalty based methods are used. So I don't know if the algoritmh may change depending of the model contact used by the simulator, for contacts Bullet use impulses not forces.
Can I use PD controllers instead of compute the correct torque to get the desired joint angles taking into account the ground contact forces? Exist a good alternative to avoid use inverse dynamics and get good results?
Thanks in advance and sorry for my bad english
