I'm trying to generate a simulation that automatically executes grasping multiple times to collect success/fail of the specific grasp configuration. Currently, I have robot moveit package (rviz&gazebo), STL files of objects, and goal end-effector poses with respect to the objects' coordinate systems.
At the moment, my primary goal is to spawn an object and a robot simultaneously to use tf (from robot base to object) to define the goal pose of an end-effector for gazebo simulation.
My initial idea of how to setup the simulation is to convert the object STL file to the URDF and load it as part of the robot in both rviz and gazebo. Is there any better idea to achieve the primary goal? I appreciate any comment, suggestion, and information.
Post score: 0