Diseño y simulación de un robot seguidor de personas para el transporte de suministros en hospitales
Files
Date
Subject
odometry
local path planning
aruco marker
global coordinates
people follow-ing robot
odometría
planeamiento de la ruta local
marcador aruco
coordenadas globales
robot seguidor de persona
local path planning
aruco marker
global coordinates
people follow-ing robot
odometría
planeamiento de la ruta local
marcador aruco
coordenadas globales
robot seguidor de persona
Language:
Journal Title
Journal ISSN
Volume Title
Publisher
Instituto Tecnológico de Santo Domingo (INTEC)
Las enfermeras suelen realizar múltiples tareas de manera simultánea en hospitales; las de medicación son las más frecuentes y exigen el transporte de equipos pesados, lo cual genera agotamiento y provoca una disminución en la calidad de su trabajo . El objetivo de este artículo es desarrollar un prototipo simulado de un robot capaz de seguir enfermeras durante su recorrido en una superficie plana, aligerando su carga de trabajo. Este robot debe ser capaz no solo de seguir a una persona en particular, identificada por medio de un marcador aruco en la espalda, sino tener la capacidad de obtener información del ambiente que le rodea mientras sigue a la persona para actuar ante los cambios ambientales y seguirla, incluso cuando desaparece de su zona de visión momentáneamente al doblar en esquinas. La entrada del sistema es la imagen del marcador, que mediante el algoritmo de identificación computa las coordenadas del centro en pixeles y distancia al robot; si la persona está presente se utiliza un controlador P y PD para el seguimiento. En caso de que el usuario desaparezca de la cámara se realiza un planeamiento de la ruta local hacia la última coordenada global extraída de la persona, mediante la cámara monocular y luego un sistema de búsqueda. Los parámetros de planeamiento se modifican para pasar a través de una puerta; dicho reconocimiento se logra mediante la red convolucional YOLO v3. La localización del robot se determina con la odometría de rueda y al doblar el usuario en una esquina se determina la dirección en la cual se estaba moviendo para localizarlo nuevamente. Para analizar el desempeño del sistema propuesto se utilizó el software CoppeliaSim con el modelo de robot Pioneer 3-DX, donde se llevan a cabo diversas simulaciones bajo distintos escenarios a los cuales estaría expuesto el robot.
Nurses tend to perform multiple tasks simultaneously, the medication tasks being the most frequent where they transport heavy carts, which usually generates ex-haustion and causes a decrease in the quality of their work. The objective of this article is to develop a simulated prototype of a robot capable of following nurses during their journey on a flat surface, lightening their workload. This robot must be capable not only of following a particular person identified by means of an aruco marker on the back, but also have the ability to obtain information from the environment around them while following the person to act on environmental changes, and following the target even when disappearing from his vision zone momentarily when turning at different corners. The input of the system is the im-age of the marker, the coordinates of the center in pixels and distance to the ro-bot are computed by the identification algorithm. If the person is present, a P and PD controller is used for following. In case the user disappears for some reason from the camera, the robot makes a local path planning to the last global coordinate of the person extracted using the monocular camera and then a search system. The planning parameters are modified to pass through a door, this recognition is achieved through the YOLO v3 convolutional network. The localization of the robot is determined with the wheel odometry and when the target user gets around corners, the direction in which it was moving is deter-mined in order to follow it again. To analyze the performance of the proposed system, the CoppeliaSim software was used with the Pioneer 3-DX robot model, where various simulations are carried out under different scenarios to which the robot would actually be exposed.
Nurses tend to perform multiple tasks simultaneously, the medication tasks being the most frequent where they transport heavy carts, which usually generates ex-haustion and causes a decrease in the quality of their work. The objective of this article is to develop a simulated prototype of a robot capable of following nurses during their journey on a flat surface, lightening their workload. This robot must be capable not only of following a particular person identified by means of an aruco marker on the back, but also have the ability to obtain information from the environment around them while following the person to act on environmental changes, and following the target even when disappearing from his vision zone momentarily when turning at different corners. The input of the system is the im-age of the marker, the coordinates of the center in pixels and distance to the ro-bot are computed by the identification algorithm. If the person is present, a P and PD controller is used for following. In case the user disappears for some reason from the camera, the robot makes a local path planning to the last global coordinate of the person extracted using the monocular camera and then a search system. The planning parameters are modified to pass through a door, this recognition is achieved through the YOLO v3 convolutional network. The localization of the robot is determined with the wheel odometry and when the target user gets around corners, the direction in which it was moving is deter-mined in order to follow it again. To analyze the performance of the proposed system, the CoppeliaSim software was used with the Pioneer 3-DX robot model, where various simulations are carried out under different scenarios to which the robot would actually be exposed.
Description
Type
info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
info:eu-repo/semantics/publishedVersion
Source
Science, Engineering and Applications; Vol. 3 No. 2 (2020): Science, Engineering and Applications; 91-126
Ciencia, Ingenierías y Aplicaciones; Vol. 3 Núm. 2 (2020): Ciencia, Ingenierías y Aplicaciones; 91-126
2636-2171
2636-218X
10.22206/cyap.2020.v3i2
Ciencia, Ingenierías y Aplicaciones; Vol. 3 Núm. 2 (2020): Ciencia, Ingenierías y Aplicaciones; 91-126
2636-2171
2636-218X
10.22206/cyap.2020.v3i2