Artificial Vision

4 DoF Robotic Arm Position Control through Artificial Vision

Objective: Develop an inverse kinematics algorithm for a 4 DoF robotic arm and a digital image processing method to identify and locate a ball of a certain color in space, so the robot can actually recognize ball’s coordinates and catch it if it’s in a reachable position.

Summary:

  • Get robotic arm kinematics diagram.
  • Develop and validate direct kinematics model.
  • Develop and validate inverse kinematics model.
  • Perform camera’s calibration.
  • Perform digital image processing from cameras in order to recognize the ball and run a stereoscopic vision algorithm in order to get ball’s centroid coordinates.
  • Calculate the distance from the robotic arm inertial frame to ball’s centroid coordinate in order to check if it’s inside robot’s work space.
  • Develop a kinematic trajectory planning in order to catch the ball with the robotic arm.

Results:

Loading file
  • Fullscreen
  • Wireframe
  • Zoom In
  • Zoom Out
  • Rotation
  • Screenshot
  • Help

Controls

×
  • Rotate with the left mouse button.
  • Zoom with the scroll button.
  • Adjust camera position with the right mouse button.
  • Double-click to enter the fullscreen mode.
  • On mobile devices swipe to rotate.
  • On mobile devices pinch two fingers together or apart to adjust zoom.
  • On mobile devices 3 finger horizontal swipe performs panning.