Dual-arm robotic manipulation represents a promising solution for the automation of complex tasks that require coordination, precision, and controlled interaction with the environment. One of the most relevant applications is automated welding, where it is necessary to follow constrained trajectories while constantly maintaining controlled contact with the workpiece. In this thesis, a dual-arm system is developed and simulated to mimic automated welding operations through precise trajectory tracking. The system is modeled in MATLAB, integrating CasADi for symbolic and numerical optimization and the Robotics System Toolbox for robot modeling and simulation. Reference trajectories are generated during the path planning phase using the Orientation-Constrained Rapidly exploring Random Tree (OC-RRT) algorithm, which enables the creation of kinematically feasible and optimized paths. Trajectory execution is handled by a 12-degree-of-freedom Nonlinear Model Predictive Controller (NMPC), designed to ensure accuracy and stability even under non-ideal conditions. The system is tested in simulation under dynamic disturbances, sensor noise, and in addition with a distance constraint between the end-effectors of the two manipulators. A comprehensive description of the system architecture, planning strategies, and control techniques is provided, along with a performance analysis in both nominal and perturbed scenarios. The results highlight the robustness and effectiveness of the proposed approach, paving the way for potential future implementations on real robotic manipulators in advanced industrial environments.
La manipolazione robotica a doppio braccio rappresenta una soluzione promettente per l' automazione di compiti complessi che richiedono coordinazione, precisione e controllo dell’ interazione con l’ ambiente. Una delle applicazioni più rilevanti è la saldatura automatica, in cui è necessario seguire traiettorie vincolate mantenendo costantemente un contatto controllato con il pezzo da lavorare. In questa tesi viene sviluppato e simulato un sistema a doppio braccio, progettato per riprodurre operazioni di saldatura automatizzata tramite un preciso tracciamento della traiettoria. Il sistema è modellato in ambiente MATLAB, integrando CasADi per l’ ottimizzazione simbolica e numerica e la Robotics System Toolbox per la modellazione e simulazione dei robot manipolatori. Le traiettorie di riferimento vengono generate nella fase di path planning tramite l’ algoritmo OC-RRT (Orientation - Constrained Rapidly exploring Random Tree), che consente la creazione di percorsi cinematicamente validi e ottimizzati. L’ esecuzione delle traiettorie è affidata a un controllore NMPC (Nonlinear Model Predictive Control) a 12 gradi di libertà, progettato per garantire precisione e stabilità anche in condizioni non ideali. Il sistema viene testato in simulazione in presenza di disturbi dinamici, rumore sui sensori e con l’introduzione di un vincolo di distanza tra gli end-effector dei due manipolatori. Si fornisce una descrizione completa dell’ architettura del sistema, delle strategie di pianificazione e delle tecniche di controllo adottate, accompagnata da un’ analisi delle prestazioni in scenari sia nominali che perturbati. I risultati evidenziano la robustezza e l’ efficacia dell’ approccio proposto, ponendo le basi per futuri sviluppi su manipolatori robotici reali in contesti industriali avanzati.
Trajectory Execution for Coordinated Dual-Arm Robotic Systems in Simulation
GARBO, ALESSIA
2024/2025
Abstract
Dual-arm robotic manipulation represents a promising solution for the automation of complex tasks that require coordination, precision, and controlled interaction with the environment. One of the most relevant applications is automated welding, where it is necessary to follow constrained trajectories while constantly maintaining controlled contact with the workpiece. In this thesis, a dual-arm system is developed and simulated to mimic automated welding operations through precise trajectory tracking. The system is modeled in MATLAB, integrating CasADi for symbolic and numerical optimization and the Robotics System Toolbox for robot modeling and simulation. Reference trajectories are generated during the path planning phase using the Orientation-Constrained Rapidly exploring Random Tree (OC-RRT) algorithm, which enables the creation of kinematically feasible and optimized paths. Trajectory execution is handled by a 12-degree-of-freedom Nonlinear Model Predictive Controller (NMPC), designed to ensure accuracy and stability even under non-ideal conditions. The system is tested in simulation under dynamic disturbances, sensor noise, and in addition with a distance constraint between the end-effectors of the two manipulators. A comprehensive description of the system architecture, planning strategies, and control techniques is provided, along with a performance analysis in both nominal and perturbed scenarios. The results highlight the robustness and effectiveness of the proposed approach, paving the way for potential future implementations on real robotic manipulators in advanced industrial environments.| File | Dimensione | Formato | |
|---|---|---|---|
|
Garbo_Alessia.pdf
accesso aperto
Dimensione
10.58 MB
Formato
Adobe PDF
|
10.58 MB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/93370