autodock

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 lib
• autodock_examples: Examples for autodock
• autodock_sim: Autodock Simulation

Installation

ROS Noetic is used during development.

Dependencies

• Fiducial
• Turtlebot Simulation ** Optional

# First, clone repos and deps to 'catkin_ws/src', then install
cd catkin_ws
rosdep update && rosdep install --from-paths src --ignore-src -yr
catkin_make

Architecture Diagram

State Machine Diagram

---

Run Examples on Simulation with MockRobot

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 a std_srvs/Trigger to the MockCharger 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

More Tests

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

---

Run with turtlebot3

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

---

Run Turtlebot3 with navigation

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 a burger 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 only pause if it is in predock, steer_dock, or parralel_correction state.

---

Docker Container

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"

---

Notes

The initial development of this code was graciously supported by Kabam Robotics (aka Cognicept).

Future Work

• Advance ObstacleObserver: filter occupancy of charging station on costmap.

Stray Commands

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