Fix the joint ordering in Simple #14
Conversation
There was a problem hiding this comment.
Hi @edantec , thanks for fixing the joint ordering.
Your PR introduces many other (unrelated) changes. Some of them refer to older versions of Simple and have either no effect or are not necessary anymore. I think we should revert them or discuss what we why you want to integrate them.
| self.simulator.contact_solver_info = pin.ProximalSettings( | ||
| args["tol"], args["tol_rel"], args["mu_prox"], args["maxit"] | ||
| ) |
There was a problem hiding this comment.
contact_solver_info does not exist anymore, please revert this change. in my PR I changed these things on purpose to make it work with the latest Simple API.
| self.simulator.constraints_problem.Kp = args["Kp"] | ||
| self.simulator.constraints_problem.Kd = args["Kd"] |
| @@ -70,6 +61,9 @@ def __init__(self, model, geom_model, visual_model, q0, args): | |||
| self.v = np.zeros(self.model.nv) | |||
| self.a = np.zeros(self.model.nv) | |||
| self.f_feet = np.zeros(4) | |||
| self.fext = [pin.Force(np.zeros(6)) for _ in range(self.model.njoints)] | |||
There was a problem hiding this comment.
i dont understand, why do you introduce zeros forces here ? this is not necessary anymore in simple step, please remove
| self.simulator.step(self.q, self.v, tau, self.fext, self.dt) | ||
| else: | ||
| self.simulator.stepPGS(self.q, self.v, tau, self.dt) | ||
| #print(self.simulator.getStepCPUTimes().user) | ||
| self.simulator.stepPGS(self.q, self.v, tau, self.fext, self.dt) | ||
|
|
| 'tol': node.declare_parameter('tol', 1e-6).value, | ||
| 'tol_rel': node.declare_parameter('tol_rel', 1e-6).value, | ||
| 'mu_prox': node.declare_parameter('mu_prox', 1e-4).value, | ||
| 'maxit': node.declare_parameter('maxit', 100).value, | ||
| 'warm_start': node.declare_parameter('warm_start', 1).value, | ||
| 'contact_solver': node.declare_parameter('contact_solver', 'ADMM').value, | ||
| 'contact_solver': node.declare_parameter('contact_solver', 'PGS').value, |
There was a problem hiding this comment.
why do you change the default solver ?
This PR fixes the joint ordering in the Simple wrapper and correctly set up the Simple solver by using PGS rather than ADMM.