Object Detection: Pick and Place Robot

This paper details the implementation of kinematic modeling, vision-based object detection, and control strategies for a 3-degree-of-freedom (3DOF) robotic arm. The project integrates camera vision and robot kinematics to identify, sort, and manipulate objects based on color and size. A MATLAB-based approach was used to calibrate the camera, process images, and transform object coordinates from the camera's reference frame to the robot's workspace.

The robotic arm's movement was controlled using forward and inverse kinematics, Jacobian calculations, and trajectory planning with quintic polynomial interpolation. PID control was implemented for precise movement of the arm’s joints, and singularity detection was incorporated to prevent unstable configurations. Additionally, force calculations were applied to maintain equilibrium under external forces.

Experimental results showed accurate object detection and successful object manipulation, with minimal transformation and positioning errors. The study highlights the importance of Jacobians for velocity control, manipulability analysis, and singularity prevention. The project provides a structured framework for integrating computer vision, kinematics, and control algorithms, demonstrating its potential for automated sorting and manipulation tasks in robotic applications.