Open jangmingyu opened 2 years ago
Can you simulate it without training? You have to first check if the URDF is working
Thanks for your reply. Again, sorry for the late reply. I've been proceeding as you said, but I'm having more difficulties than I thought, so I'm going to proceed in a different way. So I would like to ask if there is any possibility that the contents will come out differently.
1. Install linear rails on both sides of "animal_model", fix them with "animal_model", and then move the linear rails according to the movement of "animal_model" to implement movement in the z-axis direction.
2. In order to proceed with the contents of No. 1, it seems that the linear rail must operate passively with "anymal_model" operation without special control input. However, since raisim knows that each joint has a torque value, it knows that it implements linear motion by applying separate control torque values to the linear rails. But I want to implement linear motion with the power of "anymal_model". So, I searched for data and found that if I set the "Effort" of the linear rail joint to 0 in URDF, the linear rail is passively acted upon by the motion or force of "anymal_model".
I would like to know if the above method solves the problem and if there is a passive joint implementation in raism. Any additional comments to help with this would be greatly appreciated.
Hello, I am trying to limit y by connecting the base of anymal to the world and again connecting the base to base_inertia to limit movement on the y-axis when learning anymal. The world and base are currently connected as Fixed, but we plan to connect them with prismatic in the future. Below is my current URDF.
[problem]
When I try to train through that URDF, I get an error. If you have any small comments, we would appreciate it.