Autonomous Docking for ROS Robots

This repository provides ROS packages designed for precise and reliable autonomous docking of differential-drive robots. Utilizing a state machine-based approach and fiducial markers for navigation, robots equipped with camera sensors can achieve accurate docking without modifying existing navigation stacks.

---

Packages

• autodock_core: Core docking logic and libraries
• autodock_examples: Demonstrations and test scripts
• autodock_sim: Simulation environments for docking scenarios

Repository Workflow

---

Installation

Developed for ROS Noetic. Install dependencies first:

cd ~/catkin_ws/src
git clone https://github.com/ArghyaChatterjee/autonomous-docking-for-mobile-robots.git autodock
cd ~/catkin_ws
rosdep update && rosdep install --from-paths src --ignore-src -yr
catkin_make

Dependencies:

• Fiducial Marker Detection
• Turtlebot Simulation (optional)

---

Diagrams

Architecture

State Machine

---

Running Simulations

MockRobot Example in Gazebo

Launch Gazebo simulation:

roslaunch autodock_sim dock_sim.launch

Send docking goal:

rostopic pub /autodock_action/goal autodock_core/AutoDockingActionGoal {} --once

The robot initiates docking upon receiving a goal and interacts with a simulated charger (MockCharger).

Control Commands:

# Remote control robot
rosrun teleop_twist_keyboard teleop_twist_keyboard.py

# Cancel docking action
rostopic pub /autodock_action/cancel actionlib_msgs/GoalID {} --once

# Pause docking action
rostopic pub /pause_dock std_msgs/Bool True --once

Run smoke tests with random docking attempts:

rosrun autodock_examples dock_sim_test.py -c 10

---

Turtlebot3 Docking Demonstrations

Simple Docking

Install dependency:

sudo apt install ros-noetic-turtlebot3-gazebo

Launch Turtlebot3 docking:

roslaunch autodock_sim tb3_dock_sim.launch
rostopic pub /autodock_action/goal autodock_core/AutoDockingActionGoal {} --once

Docking with Navigation Stack

Ensure Turtlebot3 simulation and navigation stack are installed:

sudo apt install ros-noetic-turtlebot3*

Run simulations:

1. Launch world and docking server:

roslaunch autodock_sim tb3_nav_dock_sim.launch

2. Start the navigation stack:

export TURTLEBOT3_MODEL=burger
roslaunch turtlebot3_navigation turtlebot3_navigation.launch map_file:=$HOME/map.yaml open_rviz:=0

3. Localize and navigate robot, then send docking goal:

rostopic pub /autodock_action/goal autodock_core/AutoDockingActionGoal {} --once

The system includes obstacle detection during docking.

---

Docker Deployment

Build and run Docker container:

cd catkin_ws/src/autodock
docker build -t osrf/autodock:v1 .
docker run -it --network host osrf/autodock:v1 bash -c "$COMMAND"

---

Future Work

• Improve obstacle detection integration with costmap.

---

Debugging

Launch autodock server separately:

# Terminal 1
roslaunch autodock_sim dock_sim.launch autodock_server:=false

# Terminal 2
roslaunch autodock_core autodock_server.launch \
    autodock_config:=src/autodock/autodock_examples/configs/mock_robot.yaml