Thesis Control automático del brazo robótico “Motoman-MH50" utilizando como sensor la cámara “MotoSight2D”
Loading...
Date
2026
Authors
Journal Title
Journal ISSN
Volume Title
Program
Ingeniería Civil Electrónica
Departament
Campus
Campus Casa Central Valparaíso
Abstract
En el ámbito de la automatización industrial, el uso de brazos robóticos para tareas de paletización y manipulación de piezas constituye una aplicación recurrente, la cual tradicionalmente requiere que los objetos se encuentren en posiciones predefinidas dentro del espacio de trabajo. La incorporación de sistemas de visión por computador permite aumentar la flexibilidad de estos procesos, posibilitando la detección y manipulación de objetos ubicados en posiciones variables. El presente trabajo aborda la integración de un sistema de visión por computador con un brazo robótico industrial, con el objetivo de detectar la posición de un objeto y comandar el desplazamiento del robot para su manipulación automática. El sistema implementado está compuesto por un brazo robótico Motoman MH50 y una cámara MotoSight 2D, ambos de la marca Yaskawa. La tarea de automatización desarrollada consiste en la detección de un objeto dentro del área de trabajo y su traslado hacia una posición objetivo previamente definida. Asimismo, se evalúa la factibilidad de implementar estrategias de control automático aplicadas al movimiento del robot. Los resultados obtenidos muestran que el sistema de visión permite detectar la posición del objeto con un nivel de precisión adecuado para aplicaciones de manipulación industrial. No obstante, se identifica una imprecisión mecánica en una de las articulaciones del brazo robótico, la cual afecta la exactitud del posicionamiento final. Adicionalmente, la arquitectura del sistema no permite la implementación de un control automático en lazo cerrado, debido a la imposibilidad de realizar mediciones continuas mientras el robot se encuentra en movimiento. Como alternativa, se propone una solución basada en interrupciones, la cual permite reaccionar ante desplazamientos del objeto detectado durante la operación. Esta estrategia permitió detener correctamente el robot y reiniciar el ciclo de detección de manera controlada. En conclusión, el sistema desarrollado permite realizar aplicaciones de paletización y manipulación de piezas de forma precisa y confiable, validando la integración entre el sistema de visión por computador y el brazo robótico industrial. Como trabajo futuro, se propone habilitar la activación automática de interrupciones y estudiar la factibilidad de implementar control en tiempo real del robot, con el fin de desarrollar estrategias de control automático más avanzadas.
In industrial automation, a recurring task is the use of robotic manipulators for palletizing and object handling operations, which typically require objects to be located at predefined positions within the workspace. The incorporation of computer vision systems increases the flexibility of these processes by enabling the detection of objects at variable positions. This work addresses the integration of a computer vision system with an industrial robotic arm in order to detect the position of an object and command the robot motion for its manipulation. The system is composed of a Motoman MH50 robotic arm and a MotoSight2D camera, both manufactured by Yaskawa. An automation task is implemented in which an object is detected within the workspace and transported to a target position. Additionally, the feasibility of implementing automatic control strategies for the robot motion is evaluated. The results show that the vision system is capable of detecting the object position with good accuracy. However, a mechanical inaccuracy was identified in one of the robot joints, which affects the precision of the final positioning. Furthermore, the system architecture does not allow the implementation of closed-loop automatic control, due to the inability to perform continuous measurements while the robot is in motion. As an alternative, an interruption-based solution is proposed to react to object displacement during the robot motion. This solution successfully allowed the robot to stop and restart the detection cycle. In conclusion, the system enables precise and reliable palletizing and object manipulation applications, validating the integration between the vision system and the robotic arm. As future work, the automatic activation of interruptions is proposed, along with the study of real-time robot control to enable the implementation of advanced automatic control strategies.
In industrial automation, a recurring task is the use of robotic manipulators for palletizing and object handling operations, which typically require objects to be located at predefined positions within the workspace. The incorporation of computer vision systems increases the flexibility of these processes by enabling the detection of objects at variable positions. This work addresses the integration of a computer vision system with an industrial robotic arm in order to detect the position of an object and command the robot motion for its manipulation. The system is composed of a Motoman MH50 robotic arm and a MotoSight2D camera, both manufactured by Yaskawa. An automation task is implemented in which an object is detected within the workspace and transported to a target position. Additionally, the feasibility of implementing automatic control strategies for the robot motion is evaluated. The results show that the vision system is capable of detecting the object position with good accuracy. However, a mechanical inaccuracy was identified in one of the robot joints, which affects the precision of the final positioning. Furthermore, the system architecture does not allow the implementation of closed-loop automatic control, due to the inability to perform continuous measurements while the robot is in motion. As an alternative, an interruption-based solution is proposed to react to object displacement during the robot motion. This solution successfully allowed the robot to stop and restart the detection cycle. In conclusion, the system enables precise and reliable palletizing and object manipulation applications, validating the integration between the vision system and the robotic arm. As future work, the automatic activation of interruptions is proposed, along with the study of real-time robot control to enable the implementation of advanced automatic control strategies.
Description
Keywords
Robótica industrial, Detección de objetos, Visión por computador, Automatización
