The activity presented in this PhD thesis deal with the study and implementation of safe techniques for human-robot collaboration, with particular attention to obstacles avoidance strategies. In the first part of my PhD, the research activity has been focused on the state of the art of safety criteria, safety strategies used human-robot collaboration and obstacle avoidance strategies for collaborative robots. Starting from the case of the 3R manipulator in 2D environment, the obstacle avoidance strategy has been extended to the KUKA LBR iiwa with 7 degrees of freedom and the Universal Robot UR5 to demonstrate the adaptability and effectiveness of the proposed strategy. Several experimental tests have been carried out to verify the transferability to the real system. The second part of the thesis presents the study and implementation of vision systems for the identification of obstacles and people on the work scene. During the months that I spent at the Joanneum Research Institute (Austria), the implementation of the obstacle prevention strategy with an OptiTrack X22 motion capture system has been first addressed. In addition, a vision system based on the fusion of data from three Intel realsense D455 is introduced and studied. The goal is to identify the human skeleton in order to enable a secure collaborative application between human and robot based on machine learning. In addition, a virtual reality application, developed by the SteamVR development suite with an Unity plug-in, has been created in order to model the kinematics of the KUKA robot. The final part of the thesis deals with ergonomics and safety assessment for collaborative human-robot applications. On the basis of a real case study, ergonomic analyses and risk assessments have been carried out, leading to the complete re-design of a collaborative application. Finally, once the prototype has been developed, impact tests have been performed to simulate an accidental contact between human and robot.
L’attività presentata in questa tesi di Dottorato riguarda lo studio e l’implementazione di tecniche sicure per la collaborazione uomo-robot, con particolare attenzione alle strategie di prevenzione per il contatto con gli ostacoli. Nella prima parte del mio dottorato di ricerca, l’attività di ricerca si è concentrata sullo stato dell’arte dei criteri di sicurezza, le strategie di sicurezza utilizzate la collaborazione uomo-robot e le strategie di prevenzione per evitare il contatto tra ostacoli e robot. Partendo dal caso del manipolatore 3R in ambiente 2D, la strategia di obstacle avoidance è stata estesa al KUKA LBR iiwa con 7 gradi di libertà e all'Universal Robot UR5 per dimostrarne l'adattabilità e l'efficacia. Sono stati inoltre condotti diversi test sperimentali per verificarne la trasferibilità al sistema reale. La seconda parte della tesi presenta lo studio e l'implementazione di sistemi di visione per l'individuazione di ostacoli e persone sulla scena di lavoro. Durante i mesi trascorsi presso lo Joanneum Research Institute (Austria), è stata affrontata l’implementazione della strategia di prevenzione degli ostacoli con un sistema di motion capture OptiTrack Prime X22. Sono stati condotti test sperimentali sia per UR5 che per KUKA LBR. Inoltre, viene introdotto e studiato un sistema di visione basato sulla fusione dei dati provenienti da tre Intel RealSense D455. L'obiettivo è identificare lo scheletro dell'uomo al fine di consentire una sicura applicazione collaborativa tra uomo e robot basato sul machine learning. Inoltre, un’applicazione di realtà virtuale, sviluppata dalla suite di sviluppo SteamVR con un plug-in Unity, è stata creata per modellare la cinematica del robot KUKA. L’ultima parte della tesi riguarda la valutazione del rischio e la valutazione della sicurezza per applicazioni collaborative uomo-robot. Sulla base di un caso studio reale sono state condotte analisi ergonomiche e valutazioni del rischio che hanno portato alla completa ri-progettazione di un'applicazione collaborativa. Infine, una volta sviluppato il prototipo, sono stati eseguiti test d'impatto per simulare un contato accidentale tra uomo e robot.
Design and Implementation of Safe Human-Robot Collaboration Techniques / Scoccia, Cecilia. - (2023 Mar 23).
Design and Implementation of Safe Human-Robot Collaboration Techniques
SCOCCIA, CECILIA
2023-03-23
Abstract
The activity presented in this PhD thesis deal with the study and implementation of safe techniques for human-robot collaboration, with particular attention to obstacles avoidance strategies. In the first part of my PhD, the research activity has been focused on the state of the art of safety criteria, safety strategies used human-robot collaboration and obstacle avoidance strategies for collaborative robots. Starting from the case of the 3R manipulator in 2D environment, the obstacle avoidance strategy has been extended to the KUKA LBR iiwa with 7 degrees of freedom and the Universal Robot UR5 to demonstrate the adaptability and effectiveness of the proposed strategy. Several experimental tests have been carried out to verify the transferability to the real system. The second part of the thesis presents the study and implementation of vision systems for the identification of obstacles and people on the work scene. During the months that I spent at the Joanneum Research Institute (Austria), the implementation of the obstacle prevention strategy with an OptiTrack X22 motion capture system has been first addressed. In addition, a vision system based on the fusion of data from three Intel realsense D455 is introduced and studied. The goal is to identify the human skeleton in order to enable a secure collaborative application between human and robot based on machine learning. In addition, a virtual reality application, developed by the SteamVR development suite with an Unity plug-in, has been created in order to model the kinematics of the KUKA robot. The final part of the thesis deals with ergonomics and safety assessment for collaborative human-robot applications. On the basis of a real case study, ergonomic analyses and risk assessments have been carried out, leading to the complete re-design of a collaborative application. Finally, once the prototype has been developed, impact tests have been performed to simulate an accidental contact between human and robot.File | Dimensione | Formato | |
---|---|---|---|
Tesi_Scoccia.pdf
Open Access dal 24/09/2024
Descrizione: Tesi_Scoccia
Tipologia:
Tesi di dottorato
Licenza d'uso:
Creative commons
Dimensione
16.17 MB
Formato
Adobe PDF
|
16.17 MB | Adobe PDF | Visualizza/Apri |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.