Scholarlink

The paper is devoted to provide a comprehensive survey of the currently-deployed human-robot interaction (HRI) methodologies, especially in the medical field, and a robotic system prototype integrated through discrete components. The proposed prototype comprises Raspberry Pi 5 camera, microphone, ultrasonic (US) and infrared (IR) sensors, relay module, rechargeable battery system, DC and servo motors, Raspberry Pi 5, and Arduino Nano. Based on the YOLOv5 open source frameworks, the proposed adaptive spatiotemporal robot navigation and recognition workflow includes real-time object detection, sound-driven navigation, and predictable motion planning. Upon being stimulated by an input command for movement, the offered robotic system software model is equipped with the feature of successfully moving the robot from its source position to its destination position, following the shortest path subject to the constraint of avoiding any obstacle. Our modified YOLOv5 architecture demonstrates significant temporal improvements, 1 ms per object detection, a 1000× improvement over the baseline 1 s latency. Key optimizations include dynamic database updates, in which recognized objects are flagged as “recognized” while unknown objects are labeled as “unknown” and saved for future learning.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Tactics, Techniques, and Procedures for Improving Robot Utilization in Healthcare Monitoring

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Hesham A. Sakr,
  • Mohamed Elnahas,
  • Youssef Mohamed,
  • Abdelrahman Mohamed,
  • Selim Ashraf,
  • Patrick Girgis,
  • Tomas Gaber,
  • Kareem Mohamed,
  • Ahmed Mohamed,
  • Nader Akmal,
  • Tarek Mahmoud,
  • Abdallah Fathy

摘要

The paper is devoted to provide a comprehensive survey of the currently-deployed human-robot interaction (HRI) methodologies, especially in the medical field, and a robotic system prototype integrated through discrete components. The proposed prototype comprises Raspberry Pi 5 camera, microphone, ultrasonic (US) and infrared (IR) sensors, relay module, rechargeable battery system, DC and servo motors, Raspberry Pi 5, and Arduino Nano. Based on the YOLOv5 open source frameworks, the proposed adaptive spatiotemporal robot navigation and recognition workflow includes real-time object detection, sound-driven navigation, and predictable motion planning. Upon being stimulated by an input command for movement, the offered robotic system software model is equipped with the feature of successfully moving the robot from its source position to its destination position, following the shortest path subject to the constraint of avoiding any obstacle. Our modified YOLOv5 architecture demonstrates significant temporal improvements, 1 ms per object detection, a 1000× improvement over the baseline 1 s latency. Key optimizations include dynamic database updates, in which recognized objects are flagged as “recognized” while unknown objects are labeled as “unknown” and saved for future learning.