Hi Sherm, Thanks for your explanation. I am not an expert in dynamics but from my application, I need to solve a feasible set of q, u, and udot from a potentially infeasible set of initial guess using gradient based optimizer. So I need to supply both J (partial(perr)/partial(q)) and P to the solver...

Hi Sherm, Thanks for your detailed answer. This makes sense to me. I am assuming the same logic can be applied to other types of constraints, for example, point on a surface constraint. Then why are we using velocity errors to calculate constraint matrix P? Isn't this only true when the constraint i...

Hi, I am working on an OpenSim model that involves a point constraint (which is implemented as a ball constraint in Simbody). I have a question regards to the method "calcPositionDotErrorsVirtual". Why the derivative of the constraint error is calculated as : Vec3::updAs(&pverr[0]) = v_AS - v_AC; in...

Thanks. I know q_dot = Nu. Now I realize that N may not be square matrix. Another question I have is that how Q is related to OpenSim coordinates values in "Coordinate" window. It seems like the "coordinate" in opensim follows U, the number of mobilities. Is there a simple function that takes OpenSi...

Thanks Ayman,

Another quick question I have is why the Thoracoscapular model has state.getNQ =18 while state.getNU = 17? I realize there is an extra "Q" between "scapula_winging" and "scapular_upward_rotation" but this does not quite make sense..

Kaiwen

Hi Felix, scap = myModel.getJointSet().get("scapulothoracic") returns a generic object of type "Joint" you can downcast this to the specific type of joint, since you know the type from the model file. scap_joint = ScapulothoracicJoint.safeDownCast(scap); now there will be all the ScapulothoracicJoi...

Yes. My code is basically based on that. The difference in my code is that I don't need to add any contact structure to my model. I calculate the force at velocity stage. If you can catch why my code may get a different result, please let me know.

Hello, I am trying to implement smooth Contact model https://simbody.github.io/3.7.0/classSimTK_1_1SmoothSphereHalfSpaceForce.html and compute GRF after setting the states (q and qdot) and check whether my computed value is consistent with the value reported from OpenSim HuntCrossley Contact model. ...

Thanks a lot. This is very helpful. Looks like prescribed force works similarly with prescribed controller. I will try to implement this.

Thanks for the reply. However, this body force has to be calculated from the current state, something like a contact force. Can I update this prescribed force after the system is initialized?