Implementation of the K-Nearest Neighbor Algorithm on a 4-DoF Manipulator Robot for Color-Based Object Retrieval

Abstract

This study discusses implementing the K-Nearest Neighbor (K-NN) algorithm on a 4-DoF manipulator robot to pick up objects based on color. The main objective of this study is to design and test a robot control system capable of automatically recognizing, picking up, and moving objects according to color classification. The color detection process is carried out using a camera that extracts color component values in RGB space, which are then classified using the K-NN algorithm with parameter k = 5. The classification results form the basis for the movement of the manipulator robot, which is controlled through cubic trajectory-based trajectory planning, so that the movement of each servo motor joint is smooth and coordinated. Testing was conducted 30 times with three object color categories: red, green, and blue. The experimental results showed that the system could classify colors and pick up objects with a success rate of 96%. Minor failures occurred due to lighting variations that affected color detection results. Overall, the results of this study indicate that the integration of the K-NN algorithm and cubic trajectory is effective in improving the performance of manipulator robots for color-based object recognition and picking tasks and has the potential to be applied in computer vision-based industrial automation systems.