ROS packages for automatic docking
autodock
a state machine based auto docking solution for differential-drive robot,
allows accurate and reliable docking. It utilizes 3 fiducial markers to locate the
position of the docking station. Hence, the Robot should equip with a camera input
for fiducial marker detection. Note that, this package works on top of the navigation
stack, not making any alteration to the robot's nav stack. The solution is fully
tested on multiple simulated and actual robots.
Packages:
autodock_core
: Core autodock scripts and libautodock_examples
: Examples for autodockautodock_sim
: Autodock Simulation
ROS Noetic is used during development.
Dependencies
- Fiducial
- Turtlebot Simulation ** Optional
cd ~/catkin_ws/src
git clone https://github.com/ArghyaChatterjee/autonomous-docking-for-mobile-robot-in-ros.git autodock
cd ~/catkin_ws
rosdep update && rosdep install --from-paths src --ignore-src -yr
catkin_make
Run MockRobot
in gazebo world
roslaunch autodock_sim dock_sim.launch
try to send an action goal request
rostopic pub /autodock_action/goal autodock_core/AutoDockingActionGoal {} --once
Now you will see the robot starts to execute a series of autodocking sequence.
To explain..... when an action goal
AutoDockingActionGoal
is received by the robot, the robot will start to docking sequence. When it manages to reach the targer docking station, it will send astd_srvs/Trigger
to theMockCharger
charging station, to activate the charger. The charger will validate if the robot is docked, then send a "Success" back to the robot if all goes well. Subsequently, end the entire docking action.
# To remote control the robot:
rosrun teleop_twist_keyboard teleop_twist_keyboard.py
# To cancel an active docking action:
rostopic pub /autodock_action/cancel actionlib_msgs/GoalID {} --once
# To pause an active docking action:
rostopic pub /pause_dock std_msgs/Bool True --once
This simple smoke test script will randomly move the robot to some random position in the gazebo world, and initiate the docking sequence.
# indicate the number of "spawn" with the -c arg
rosrun autodock_examples dock_sim_test.py -c 10
This demonstrates front dock with turtlebot3, attached with a front camera. A smaller docking station is used here.
- dependency:
sudo apt install ros-noetic-turtlebot3-gazebo
roslaunch autodock_sim tb3_dock_sim.launch
rostopic pub /autodock_action/goal autodock_core/AutoDockingActionGoal {} --once
This demo depends on turtlebot3_simulation and move_base. Please ensures that you have all dependencies are fully installed.
Note: You can install all turtlebot simulation related packages by
sudo apt install ros-noetic-turtlebot3*
, or follow the setup steps in Robotis docs. Once done, continue to follow the steps to launch and map the world with aburger
robot.
Please ensures that entire pipeline of the turtlebot works before proceeding with the
autodock tb3 demo below. Once done with mapping, you can start with the autodock
simulation:
# 1. launch turtlebot3 gazebo world, and also autodock_server
roslaunch autodock_sim tb3_nav_dock_sim.launch
# 2. launch navigation stack, with provided map
## New Terminal
export TURTLEBOT3_MODEL=burger
roslaunch turtlebot3_navigation turtlebot3_navigation.launch map_file:=$HOME/map.yaml open_rviz:=0
# 3. Now, localize the robot and move the robot around the world.
# 4. send a dock action. make sure that the robot's camera is facing the charger
## New Terminal
rostopic pub /autodock_action/goal autodock_core/AutoDockingActionGoal {} --once
This will also demonstrate how the simple obstacle_observer
works.
Try place an obstacle near the robot during docking and see if it pauses.
Note: the current
obstacle_observer
will onlypause
if it is inpredock
,steer_dock
, orparralel_correction
state.
A docker file is provieded to buid a container for autodock. Follow these steps:
cd catkin_ws/src/autodock
docker build -t osrf/autodock:v1 .
docker run -it --network host osrf/autodock:v1 bash -c "$COMMAND"
The initial development of this code was graciously supported by Kabam Robotics (aka Cognicept).
- Advance
ObstacleObserver
: filter occupancy of charging station on costmap.
Debug by launching autodock node on a seperate terminal
# Terminal 1
roslaunch autodock_sim dock_sim.launch autodock_server:=false
# Terminal 2, specify config.yaml with the 'autodock_config' arg
roslaunch autodock_core autodock_server.launch \
autodock_config:=src/autodock/autodock_examples/configs/mock_robot.yaml