2016 Manipulation Status
Status on March 5, 2016
- Failed to determine feasibility of our simple pendulum arm design via simulation in Gazebo. i.e. Do we need additional joints? What is optimal arm length? -- Too much time was spent trying to model our base from scratch (some simple flaw not obvious to me). We can use one of the predefined robot models in Gazebo for the base till we figure it out.
- But we do have a first cut at a software architecture for manipulation, including a high-level joint controller in Python. We are using PID controllers in the Control package for joint manipulation. These controllers receive/send ROS messages for position commands and state.
Goals for April 2, 2016
- Have the gripper functioning from high-level software.
- Have code to direct gripper positioning to pick up the can. Have it running in simulation. Initial plan is to use camera data to detect can position.
- Have Arm v 0.1 fabricated (sans sensors)