Yes. You are right Dimitar. I was mistakenly using the CoordinateSet in the second argument of the method instead of the CoordinateReference. Thanks for the pointer.
Musa

Declaring assemblysolver object: I read an open-link model from an XML file (skeleton_file.osim). I then added bodies to it so that it will become a closed chain body system after assembly. A snippet of my code is attached. snippet_code.cpp 'weights' vector is to ensure that some of the joint angles...

In the C++ code of ExampleClosedTopologyMechanism of Simbody, a ball constraint was defined using the form: "Constraint::Ball(linker1, Vec3(linkerHalfDims[0],0,0), rocker1, Vec3(0,-rockerHalfDims[1],0));" When I used the same form in my code as: SimTK::Constraint::Ball(parent_body, loc_p, child_body...

Hi, My model file works well in 3.3. However, when I tried to run a simulation with it in 4.0 C++ API, it threw an exception with the following message: --------------------------------------------------------------------------------------- Updating Model file from 30516 to latest format... Loaded m...

I found the solution. The sample Simbody CMake file says to add "include_directories(${Simbody_INCLUDE_DIR})"; and says this is "...only necessary if Simbody was built using CMake older than 3.0.2". I am using CMake version 3.12.1 and I only got the thing to work by adding that line! Thanks.
Musa

Hi, I successfully installed Simbody 3.6 and OpenSim 4.0 on Centos Linux and all tests passed for both. I installed Simbody in /usr/local/ and OpenSim in /opt/opensim-core/ (with Simbody copied to /opensim-core directory). When I tried to build an example problem, it always fails with an error of fo...

Chris, Oops! Scratch that off! I had thought I needed to override those methods when I was getting some weird error messages; but got around them. Now I did not need to override them. Thanks for all the insight. Now my code runs well. By the way, is version 4.0 now considered a stable release or is ...

Thanks Chris. This is helpful to know. I guess I can run my integration in smaller time steps (say 0.01) and call getControls() at the end of each time step and record the control values. The only thing is I will then lose the ability to use printControlStorage( ) and getStateStorage().print which a...

Thanks Chris. That works. But doesn't adInControls() add to the values already occupying the slot in the system-wide model controls? I just overrode Actuator_ setControls() in my Actuator methods to update my actuator control values and it seems to work fine. Thanks for the help.
Musa

I have a C++ model that has m muscles and n user-actuators. Each muscle has one control variable and each actuator has 3 control variables - for a total model control vector of length m+3n. The length and content of my model controls does contain the controls for all muscles and actuators in the mod...