The world of industrial robotics has been trying in recent years to enhance two aspects concerning robots, namely autonomy and flexibility. Robots with these characteristics are capable of sensing the world around them, responding to unanticipated events, and adapting to different operational situations without the need for a human intervention. The first steps in this direction were taken with the advent of collaborative robots, which introduced several advantages over classical industrial robots but which without the support of sensors or 3D cameras and intelligent algorithms still cannot have high degrees of autonomy and flexibility. Therefore, the following work proposes a possible solution to enhance these two aspects by integrating a 3D vision system and an obstacle avoidance algorithm that enable a pick and place operation with objects of different sizes and avoidance of obstacles that occupy the robot's ideal trajectory. The proposed application was developed in the ROS2 environment, which provides a robust platform for the design of complex robotic systems, integrating image processing libraries for camera management. Finally, a possible preliminary solution preparatory to the future development of this work is proposed, which is the use of a probabilistic 3D map called Octomap to also implement dynamic obstacle avoidance algorithms capable of avoiding even moving obstacles.
Il mondo della robotica industriale sta cercando negli ultimi anni di potenziare due aspetti riguardanti i robot, ovvero l’autonomia e la flessibilità. Robot con queste caratteristiche sono in grado di percepire il mondo che li circondano, di rispondere ad eventuali eventi non previsti e di adattarsi a differenti situazioni operative senza che sia necessaria la mano dell’uomo. I primi passi in questo senso sono stati fatti con l’avvento dei robot collaborativi, i quali hanno introdotto diversi vantaggi rispetto ai classici robot industriali ma che senza il supporto di sensori o camere 3D e di algoritmi intelligenti non possono comunque avere dei gradi elevati di autonomia e flessibilità. Il seguente lavoro propone quindi una possibile soluzione per incrementare questi due aspetti, integrando un sistema di visione 3D ed un algoritmo di obstacle avoidance che consentono di effettuare una operazione di pick and place con oggetti di dimensioni differenti e di evitare, se fisicamente possibile, gli ostacoli che occupano la traiettoria ideale del robot. L’applicazione proposta è stata sviluppata in ambiente ROS2, che offre una piattaforma robusta per la progettazione di sistemi robotici complessi, integrando librerie di elaborazione immagini per la gestione della camera. Si propone infine una possibile soluzione preliminare propedeutica allo sviluppo futuro di questo lavoro, che consiste nell’utilizzo di una mappa 3D probabilistica definita Octomap per implementare anche algoritmi di obstacle avoidance dinamici, in grado di evitare anche ostacoli mobili.
Integrazione di visione 3D e obstacle avoidance in ambiente ROS2 per incrementare l’autonomia di un robot collaborativo
GAUDENI, FRANCESCO
2022/2023
Abstract
The world of industrial robotics has been trying in recent years to enhance two aspects concerning robots, namely autonomy and flexibility. Robots with these characteristics are capable of sensing the world around them, responding to unanticipated events, and adapting to different operational situations without the need for a human intervention. The first steps in this direction were taken with the advent of collaborative robots, which introduced several advantages over classical industrial robots but which without the support of sensors or 3D cameras and intelligent algorithms still cannot have high degrees of autonomy and flexibility. Therefore, the following work proposes a possible solution to enhance these two aspects by integrating a 3D vision system and an obstacle avoidance algorithm that enable a pick and place operation with objects of different sizes and avoidance of obstacles that occupy the robot's ideal trajectory. The proposed application was developed in the ROS2 environment, which provides a robust platform for the design of complex robotic systems, integrating image processing libraries for camera management. Finally, a possible preliminary solution preparatory to the future development of this work is proposed, which is the use of a probabilistic 3D map called Octomap to also implement dynamic obstacle avoidance algorithms capable of avoiding even moving obstacles.File | Dimensione | Formato | |
---|---|---|---|
Gaudeni_TesiMagistrale.pdf
non disponibili
Descrizione: Caricamento tesi magistrale Gaudeni
Dimensione
34.45 MB
Formato
Adobe PDF
|
34.45 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.12075/13680