4-DoF Robotic Arm for Picking and Moving RGB Color-Based Objects Using the Support Vector Machine Method

Abstract

This study discusses designing and implementing an RGB color pattern recognition system using the Support Vector Machine (SVM) method on a 4-DoF robotic arm to perform autonomous object transfer tasks. This system integrates computer vision, artificial intelligence, and trajectory planning technologies to improve the adaptability and precision of the robot manipulator's movements. The pattern recognition process is done through image acquisition using a camera mounted on a support pole, then extraction and normalizing color values in the R, G, and B channels. These RGB values are input features for color pattern classification using SVM with Radial Basis Function (RBF) kernel and regulation parameter C = 100. The training results show that the SVM model can classify three color classes (red, yellow, and blue) with an accuracy rate of 100%. The classification data is then used to control the movements of three robots with red, orange, and blue arms, each tasked with picking up and moving objects of the corresponding color. The robot trajectory was planned using the Cubic Trajectory method, which produced smooth and coordinated movements between joints, with an average task completion time of ±10 seconds. Based on the results of 30 trials, the system showed a success rate of 96.67%, with only one failure due to gripper position inaccuracy. The results of this study indicate that the combination of the SVM and Cubic Trajectory methods can improve the efficiency and accuracy of robotic arm systems in color-based object recognition and manipulation, which has the potential to be applied to artificial intelligence-based industrial automation systems.