Wira Aditama, David Herianto, Nico Fernando, Carolus Henry, Nova Eka Budiyanta

Design and Implementation of a Vision-based Wheeled Mobile Robot Using HSV Color Segmentation and P-D Control

Introduction

Design and implementation of a vision-based wheeled mobile robot using hsv color segmentation and p-d control. Design and implement a vision-based wheeled mobile robot using HSV color segmentation and P-D control. Tracks a ping-pong ball with Raspberry Pi, demonstrating feasibility for simple object tracking.

Abstract

This study presents the design and implementation of a wheeled mobile robot capable of detecting and tracking a ping-pong ball using vision-based processing. The system integrates a Raspberry Pi 3 Model B+ as the main controller, a Raspberry Pi Camera Rev 1.3 for visual input, and DC motors driven by an L298N motor driver for actuation. Object detection is achieved through color segmentation in the HSV color space using the OpenCV library, followed by morphological filtering and contour analysis. A proportional-derivative (PD) control algorithm is employed to adjust motor speeds dynamically based on the ball's horizontal position in the frame. The experimental results demonstrate that the robot can successfully detect and follow a ping-pong ball, although it exhibits limitations in processing speed and motion stability. The average frame rate during operation was 5 FPS, which is sufficient for basic tracking tasks but suboptimal for high-speed applications. This project highlights the feasibility of vision based robotic systems for simple object tracking tasks.

Review

This study presents a practical implementation of a vision-based wheeled mobile robot designed for tracking a ping-pong ball, utilizing readily available components and established image processing techniques. The authors effectively outline the system's architecture, including the Raspberry Pi controller, camera, and motor drivers, along with the core methodologies of HSV color segmentation, morphological filtering, and contour analysis for object detection. The deployment of a Proportional-Derivative (PD) control algorithm to manage motor speeds based on the ball's position demonstrates a functional approach to closed-loop control in a robotic system. The work successfully validates the feasibility of simple vision-based object tracking, offering a clear and reproducible example of integrating perception and control in mobile robotics. While the project successfully demonstrates basic tracking capabilities, the abstract candidly highlights critical limitations that warrant further attention. The reported average frame rate of 5 FPS, though deemed sufficient for "basic tracking tasks," significantly restricts the robot's potential for high-speed or more dynamic applications. This low processing speed could contribute to the observed "motion stability" issues, as delayed feedback can lead to jerky movements or overshoots. Furthermore, the term "motion stability" itself is quite broad; a more detailed analysis of the instability (e.g., oscillations, tracking errors under varying speeds, or response to external disturbances) would provide valuable insight into the control system's performance and mechanical design. Despite these limitations, the paper contributes a valuable proof-of-concept for entry-level vision-based robotics. Future work should prioritize addressing the identified processing speed bottleneck, perhaps through hardware upgrades (e.g., a more powerful embedded system or dedicated vision processing unit), algorithm optimization, or exploring lighter-weight image processing techniques. Enhancing motion stability could involve advanced control strategies, sensor fusion, or mechanical improvements to the robot's chassis and drivetrain. Expanding on this foundation, the project could explore tracking multiple objects, different types of objects, or operating in more complex environments. Overall, this work serves as a solid base for further development in accessible robotic vision systems.

Full Text

Comments

You need to be logged in to post a comment.