arduino_robot_arm

ROS packages that can be used to plan and execute motion trajectories for a robot arm in simulation and real-life.

These packages were tested under ROS kinetic and Ubuntu 16.04 and it works perfectly on ROS melodic and noetic

The robot arm uses Moveit plugin to apply kinematics by the KDL solver. These packages can be tested in the gazebo simulation tool and the real robot arm, where the ROS system and Arduino code share the /joint_states topic to control motors.

Dependencies

run this instruction inside your workspace:

$ rosdep install --from-paths src --ignore-src -r -y

make sure you installed all these packages:

for kinetic distro

$ sudo apt-get install ros-kinetic-moveit
$ sudo apt-get install ros-kinetic-joint-state-publisher ros-kinetic-joint-state-publisher-gui
$ sudo apt-get install ros-kinetic-gazebo-ros-control joint-state-publisher
$ sudo apt-get install ros-kinetic-ros-controllers ros-kinetic-ros-control

for melodic distro

$ sudo apt-get install ros-melodic-moveit
$ sudo apt-get install ros-melodic-joint-state-publisher ros-melodic-joint-state-publisher-gui
$ sudo apt-get install ros-melodic-gazebo-ros-control joint-state-publisher
$ sudo apt-get install ros-melodic-ros-controllers ros-melodic-ros-control

for noetic distro

$ sudo apt-get install ros-noetic-moveit
$ sudo apt-get install ros-noetic-joint-state-publisher ros-noetic-joint-state-publisher-gui
$ sudo apt-get install ros-noetic-gazebo-ros-control joint-state-publisher
$ sudo apt-get install ros-noetic-ros-controllers ros-noetic-ros-control

Robot Arm

The robot arm has 5 joints only 4 joints can be fully controlled via ROS and Rviz, the last joint (gripper) has a default motion executed from the Arduino code directly.

Circuit diagram

Robot initial positions

Configuring Arduino with ROS

Install Arduino IDE in Ubuntu https://www.arduino.cc/en/software to install run $ sudo ./install.sh after unzipping the folder
Launch the Arduino IDE
Install the arduino package and ros library http://wiki.ros.org/rosserial_arduino/Tutorials/Arduino%20IDE%20Setup
Make sure to change the port permission before uploading the Arduino code $ sudo chmod 777 /dev/ttyUSB0

Usage

Controlling the robot arm by joint_state_publisher

$ roslaunch robot_arm_pkg check_motors.launch

You can also connect with hardware by running:

$ rosrun rosserial_python serial_node.py _port:=/dev/ttyUSB0 _baud:=115200

(Note: You may need to use ttyACM)

Simulation

Run the following instructions to use gazebo

$ roslaunch robot_arm_pkg check_motors.launch
$ roslaunch robot_arm_pkg check_motors_gazebo.launch
$ rosrun robot_arm_pkg joint_states_to_gazebo.py

(You may need to change the permission)

$ sudo chmod +x ~/catkin_ws/src/arduino_robot_arm/robot_arm_pkg/scripts/joint_states_to_gazebo.py

Controlling the robot arm by Moveit and kinematics

$ roslaunch moveit_pkg demo.launch

You can also connect with hardware by running:

$ rosrun rosserial_python serial_node.py _port:=/dev/ttyUSB0 _baud:=115200

(Note: You may need to use ttyACM)

Simulation

Run the following instruction to use gazebo

$ roslaunch moveit_pkg demo_gazebo.launch

Pick and place by using OpenCV

Preparation

Download webcam extension for VirtualBox

https://scribles.net/using-webcam-in-virtualbox-guest-os-on-windows-host/

Testing the camera and OpenCV

Run color_thresholding.py to test the camera

Before running, find the camera index normally it is video0

$ ls -l /dev | grep video

If it is not, update line 8 in color_thresholding.py

8 cap=cv2.VideoCapture(0)

Then run

$ python color_thresholding.py

Using OpenCV with the robot arm in ROS

In Real Robot

In a terminal run

$ roslaunch moveit_pkg demo.launch

this will run Rviz

connect with Arduino:

select the Arduino port to be used on Ubuntu system
change the permissions (it might be ttyACM)

$ ls -l /dev | grep ttyUSB

$ sudo chmod -R 777 /dev/ttyUSB0
upload the code from Arduino IDE

$ rosrun rosserial_python serial_node.py _port:=/dev/ttyACM0 _baud:=115200

In another terminal

$ rosrun moveit_pkg get_pose_openCV.py

This will detect blue color and publish the x,y coordinates to /direction topic

(Note: check the camera index and update the script if needed)

Open another terminal

$ rosrun moveit_pkg move_group_node

This will subscribe to /direction topic and execute motion by using Moveit move group

The pick and place actions are performed from the Arduino sketch directly.

In simulation (Gazebo)

In a terminal run

$ roslaunch moveit_pkg demo_gazebo.launch

this will run Rviz and gazebo

In another terminal

$ rosrun moveit_pkg get_pose_openCV.py

This will detect blue color and publish the x,y coordinates to /direction topic

(Note: check the camera index and update the script if needed)

Open another terminal

$ rosrun moveit_pkg move_group_node

This will subscribe to /direction topic and execute motion by using Moveit move group

We can’t visualize the pick and place actions in gazebo

wangwenqiang2019 / arduino_robot_arm Goto Github PK