This package intends to provide an implementation of an automated inspection robot, which can map its environment, plan paths, and dynamically update for obstacle avoidance. The inspection routine is a simple 'defective' check for color on 'cans' spawned at dynamically specified locations. Once specified, the robot will move to a pose slightly offset from the position to inspect the can using its built in sensors and cameras, repeat for each can, them go back to its home position.
The robot used is the TIAGo robot, chosen because of its all-in-one solution for motion and sensing, as well as its massive support for open-sourced ROS/Gazebo/RViz solutions.
This package builds on top of those utilities with its own custom controller to interface with the robot. Below describes the procedure to prepare the environment, run the main controller, and spawn a list of 'cans' for inspection.
TODO
TODO
Nick graduated from his undergraduate program in the summer of 2019 with a skillset in computer engineering, software engineering, mathematics, and physics. He was a robotics engineer at a collaborative robotics startup for two years, and has been a contractor for NASA as a robotics software engineer for the last 1.5 years. Seeking a formal education in it, he is also about halfway through the robotics engineering program at UMD, in order to strengthen his otherwise self-taught background. His hobbies include creating software, rock climbing, snowboarding, bonfires, and, more recently, 3D modeling.
See the license for this project here.
To Project Backlog (Product Backlog, Iteration Backlogs, and Work Log):
https://docs.google.com/spreadsheets/d/1DmnGjTfYCdlwXq4yxJ25zSwCLq8LcW4DftqtF5_p5Tk/edit?usp=sharing
To Sprint Planning Notes/Review:
https://docs.google.com/document/d/1JHqd9Alk2kZUKKPmb6cLGOPEvR-gZKouV5svNGUF8SE/edit?usp=sharing
Furthermore, see the "UML" directory of this package for UML files, and Proposal deliverables of the initial proposal in the "Proposal" directory.
- The controller API provided here is tightly bound to ROS and Gazebo. This may be acceptable considering the TIAGo utilities are as well, but there it'd be beneficial to explore decoupling this controller from ROS utilities.
- The CanCharacterizer class unfortunately remains unimplemented. The main controller has stubbed out its interaction with this node for now and can move as would be expected during the actual inspection routine, but no inspection is performed at the moment.
- The automated build jobs, including the tests, take a while to finish because a lot of packages are built from source. Many of these are possibly unnecessary for this package's intentions and could be removed as dependencies. This was put off for the future in the interest of time.
- The frames for the RGB camera and depth sensor are assumed to be the same, but this is not true in reality. There is a very slight, one-dimensional, translational offset that should really be accounted for.
- There could be more intensive integration tests put in place, but were put off as future work in the interest of time. Especially considering the automated builds and tests already take a while.
The following should install all of the required packages (assuming the proper sources have been declared, see ROS Melodic installation otherwise):
sudo apt-get update
sudo apt-get install git python-rosinstall ros-melodic-desktop-full python-catkin-tools ros-melodic-joint-state-controller ros-melodic-twist-mux ros-melodic-ompl ros-melodic-controller-manager ros-melodic-moveit-core ros-melodic-moveit-ros-perception ros-melodic-moveit-ros-move-group ros-melodic-moveit-kinematics ros-melodic-moveit-ros-planning-interface ros-melodic-moveit-simple-controller-manager ros-melodic-moveit-planners-ompl ros-melodic-joy ros-melodic-joy-teleop ros-melodic-teleop-tools ros-melodic-control-toolbox ros-melodic-sound-play ros-melodic-navigation ros-melodic-depthimage-to-laserscan ros-melodic-moveit-commander
- Create a catkin workspace and its src directory at e.g. /path/to/tiago_ws/src
- Clone this package into the src directory:
cd /path/to/tiago_ws/src
git clone https://github.com/rnvandemark/enpm808x_final_inspection_robot.git
- Navigate to the catkin workspace:
cd /path/to/tiago_ws/
- Use rosinstall to download additional packages into your workspace, using the additional dependencies declare in this repository's rosinstall file:
# If needed, replace 'enpm808x_final_inspection_robot' with the local name you
# gave this repository
rosinstall src /opt/ros/melodic src/enpm808x_final_inspection_robot/dependencies.rosinstall
- Set up rosdep and ensure any missing dependencies required by this workspace are met/installed:
sudo rosdep init
rosdep update
cd /path/to/tiago_ws/
rosdep install --from-paths src --ignore-src -y --rosdistro melodic --skip-keys="opencv2 opencv2-nonfree pal_laser_filters speed_limit_node sensor_to_cloud hokuyo_node libdw-dev python-graphitesend-pip python-statsd pal_filters pal_vo_server pal_usb_utils pal_pcl pal_pcl_points_throttle_and_filter pal_karto pal_local_joint_control camera_calibration_files pal_startup_msgs pal-orbbec-openni2 dummy_actuators_manager pal_local_planner gravity_compensation_controller current_limit_controller dynamic_footprint dynamixel_cpp tf_lookup opencv3 joint_impedance_trajectory_controller cartesian_impedance_controller omni_base_description omni_drive_controller"
- Optionally, add this to ~/.bashrc (and resource as necessary) to restrict message generation for extra languages:
export ROS_LANG_DISABLE=genlisp:gennodejs:geneus
- Prerequisite: you have set up your workspace as described in the previous section, assuming the catkin workspace at e.g. /path/to/tiago_ws/
- Navigate to the catkin workspace and build with the build script:
cd /path/to/tiago_ws/
./src/enpm808x_final_inspection_robot/bin/build-ws.sh
The tests for this package are automatically built with the aforementioned build script, build-ws.sh. See the 'Building the Package' section.
There are two steps to running the program. The first step, building a digital map of the environment, only has to be done once (but can be done multiple times). Then, every time the 'main' program is ran, it will use the saved map. There are launch files for each of these routines.
- Prerequisites:
- You have set up your workspace as described in the 'Setting Up Your Workspace' section, assuming the catkin workspace at e.g. /path/to/tiago_ws/
- You have built the package as described in the 'Building the Package' section
- For each terminal opened, make sure your ROS installation and this workspace are sourced, for example:
source /opt/ros/melodic/setup.bash
source /path/to/tiago_ws/install_isolated/setup.bash
First, launch the mapping configuration with the following command:
roslaunch enpm808x_final_inspection_robot create_map.launch
This will start Gazebo, RViz, and multiple TIAGo nodes/utilities. The map is initially empty, and we will have the TIAGo navigate around the environment to build it. To move the robot around, navigate to the RViz window and use the '2D Nav Goal' tool:
This can be clicked anywhere on the map that has already has been mapped (white pixels). Move around until the map is fully populated, i.e. all floors and walls/obstacles are mapped:
IN A NEW TERMINAL, run the following command to save the map you have created:
rosservice call /pal_map_manager/save_map "directory: ''"
This will save the map at ~/.pal/tiago_maps/config. Confirm it exists. Then, all terminals' processes can be stopped. Record the absolute file path that the map files are saved at, e.g. the output of the service call may look like:
success: True
name: "2021-12-10_021714"
full_path: "/home/lu18/.pal/tiago_dual_maps/"
message: "Map saved: 2021-12-10_021714"
So you should record the path as the concatenation of the full path, its name, and another 'configurations' subdirectory
/home/lu18/.pal/tiago_dual_maps/configurations/2021-12-10_021714
Prerequisite: you have created a map as described in the previous section
Launch the main program's configuration with the following command:
# Replace '/path/to/maps' with the path to the map data, the example in the
# previous section was '/home/lu18/.pal/tiago_dual_maps/2021-12-10_021714'
roslaunch enpm808x_final_inspection_robot main.launch map:="/path/to/maps"
This will start Gazebo with the specified map, RViz, multiple TIAGo nodes/utilities, and nodes created in this project. There are multiple optional arguments for this launch file:
- view_image: '1' to also display the raw RGB image in a separate window, '0' otherwise. Default value is '0'.
- robot: The TIAGo robot model to use, either 'steel' or 'titanium'. Default value is 'titanium'.
- tiago_start_pose: The pose, relative to the world/map frame, to start the TIAGo at. Default value is '-x 0.0 -y 0.0 -z 0.0 -R 0.0 -P 0.0 -Y 0.0'.
- extra_gazebo_args: Any additional arguments to also send to Gazebo. The default value is none.
Then, cans for inspection must be spawned. There is a demo launch file to spawn a fixed set of cans: 'demo.launch'. The launcher 'spawn_cans_from_list.launch' can be used to spawn a dynamically typed list. For example, this spawns three cans each with a couple of characteristics:
roslaunch enpm808x_final_inspection_robot spawn_cans_from_list.launch can_args_list:="1,1.0,1.0,0.0 : 0,2.0,1.0,0.0 : 1,-1.0,5.0,0.2"
The characteristics for each can are separated by a colon, and each individual characteristic is separated by a comma. These are as follows:
- Whether the can is nominal (1) or defective (0).
- The remaining three are the {x,y,z} coordinates of the initial spawn point, in meters, relative to the map's coordinate system.
The tests for this package can be ran by navigating to the project workspace directory and running the helper script:
cd /path/to/tiago_ws/
./src/enpm808x_final_inspection_robot/bin/run-tests.sh
There is a Doxygen configuration file in the project's root directory, .doxyfile. The output is generated in the docs/doxygen/ directory. There is already sample documentation there, but updated documentation can be generated with the following command:
doxygen .doxyfile
https://drive.google.com/file/d/1cki9b_330VcfZ-kmi1wQiNwIZLHwVnTX/view?usp=sharing