The goals of this lab are
- Introduction to the ROS’ motion planning tool OMPL
- Understand the sensor and environments models available in ROS
- Experimentally evaluate the performance of a sampling-based path planner
- Design of a challenging motion-planning environment
- Experimental evaluation of planning time and trajectory length
- 3 weeks
Deliverables are due on Monday after the last lab.
- Download the OMPL tutorial from the Kavraki lab
- Perform the exercises up to Section 7. These were the steps that were taken to import the ClamArm model into OMPL and ROS.
- Note: in section 4.3 you can skip the step that says echo “source … This part is already setup in the VM.
- Note: in section 4.5 you can skip the step that says export.. and echo… This part is already setup in the VM.
- Construct a planning environment using warehouse elements that requires the arm to reach around an object.
- Dig into to the code base to find out
- where the path is calculated
- where the path-length is calculated
5. Plan multiple paths for the same motion. Record the path length and planning time distributions.
Viewing Planner Diagnostic Info:
- Edit the launch file using:
rosed pr2_test_arm_navigation ompl_planning.launch
<param name="publish_diagnostics" type="bool" value="true" />
right after the
- Now launch the warehouse viewer as described in the tutorial using:
roslaunch pr2_test_arm_navigation planning_scene_warehouse_viewer_pr2_test.launch
- View the diagnostic info:
rostopic echo -c /ompl_planning/diagnostics