Alif Daffa Dziqy Riyansah, Febrian Hadiatna, Ratna Susana

Evaluation of YOLOv8n Performance for Real-time Human Detection on Autonomous Mobile Robots

Introduction

Evaluation of yolov8n performance for real-time human detection on autonomous mobile robots. Evaluate YOLOv8n for real-time human detection on autonomous mobile robots using edge computing. Achieves high precision, recall, and speed for surveillance and obstacle avoidance.

Abstract

This study presents the implementation and evaluation of the You Only Look Once version 8 nano (YOLOv8n) algorithm for real-time human detection on an autonomous mobile robot. The proposed system is designed as an edge-computing-based surveillance solution for monitoring restricted or difficult-to-access areas. The hardware platform integrates a Raspberry Pi 4B for visual processing and an Arduino Mega 2560 for navigation control through serial communication. Human detection is performed using a night-vision camera, while obstacle avoidance is supported by three ultrasonic sensors. A custom dataset was collected under various human postures, object distances ranging from 1 to 10 meters, and different lighting conditions. The YOLOv8n model was trained using 300 epochs with an image resolution of 640 × 640 pixels. Experimental results demonstrate that the proposed system achieves reliable real-time performance under varying environmental conditions. Under lighting variation tests, the model achieved 100% precision, 93.5% recall, 96.6% F1-score, and 93.55% accuracy with an average processing speed of 24.30 frames per second. Distance-based testing produced 100% precision, 92.42% recall, 96.06% F1-score, and 92.42% accuracy at 23.2 frames per second. Furthermore, autonomous navigation experiments confirmed that the robot was capable of simultaneously detecting humans and avoiding obstacles with response times ranging from 2.4 to 3.2 seconds. These findings indicate that You Only Look Once version 8 nano (YOLOv8n) provides an effective balance between detection accuracy, processing speed, and computational efficiency, making it suitable for deployment on edge-computing-based autonomous mobile robots.

Review

This study presents a compelling and timely evaluation of the YOLOv8n algorithm for real-time human detection on an autonomous mobile robot, targeting edge-computing-based surveillance solutions. The authors skillfully integrate a Raspberry Pi 4B for visual processing with an Arduino Mega 2560 for navigation, leveraging a night-vision camera and ultrasonic sensors to achieve a comprehensive detection and avoidance system. The chosen architecture and components are highly relevant for practical, low-cost deployments in restricted or challenging environments, making the proposed system a significant step towards autonomous safety and monitoring applications. The experimental results robustly support the claims of reliable real-time performance. The use of a custom dataset, collected under varying human postures, object distances (1-10 meters), and diverse lighting conditions, adds substantial credibility to the model's training and evaluation. The reported metrics are impressive, with the lighting variation tests yielding 100% precision and a high F1-score of 96.6% at 24.30 FPS, and distance-based tests maintaining similar high performance at 23.2 FPS. Crucially, the successful integration with autonomous navigation, demonstrated by simultaneous human detection and obstacle avoidance with acceptable response times (2.4-3.2 seconds), confirms the system's practical viability. These findings collectively underscore YOLOv8n's excellent balance of detection accuracy, processing speed, and computational efficiency on resource-constrained edge devices. While the abstract clearly highlights the system's strengths and robust performance, a deeper dive into the specifics of the custom dataset's diversity (e.g., representation of occlusions, complex backgrounds, or varying human speeds) would further solidify the claims of real-world applicability. Additionally, although the abstract positions YOLOv8n as suitable, a brief comparative analysis against other state-of-the-art lightweight object detection models or even earlier YOLO versions *running on the same edge hardware* could provide valuable context regarding its competitive advantage. Nonetheless, the presented work offers a strong foundation for deploying advanced human detection capabilities on autonomous mobile robots, demonstrating a well-engineered solution for critical surveillance and safety tasks.

Full Text

Comments

You need to be logged in to post a comment.