Pick−and−Place Robotic Arm

Abstract

This project was about designing and building an arm that can spot and sort objects, by color. The system uses a webcam to take pictures of objects on a table. Then OpenCV looks at the pictures to find their colors and guess where they are. The robotic arm was made in Autodesk Fusion. Printed in 3D using ABS parts. Its motion model used Denavit–Hartenberg parameters and a URDF file. To move the arm, we first tried numerical ways to solve inverse kinematics. In the end we used the IKPy library, which uses the Levenberg–Marquardt algorithm to calculate the angles for a desired end‑effector position. A Raspberry Pi runs the vision tasks. An ESP32 talks to the servo motors. Carries out the moves. The vision system does camera calibration cleans up the image and uses HSV color thresholding to find objects and set the pick‑and‑place spots. I tested the arm with three objects of different colors and it finished the sorting task.