Robotic rehabilitation of upper limbs represents a rapidly evolving field, where human-robot interaction plays a fundamental role in patients’ motor recovery. This thesis aims to develop and validate an impedance control system for the TIAGo rehabilitation robot, implementing a graduated assistance strategy based on force fields. The proposed system is based on the Assist-As-Needed paradigm, where the robot maintains a predominantly passive behavior, leaving the patient as the primary driver of movement. Assistance is provided gradually through a virtual force field constructed around the reference trajectory, modulating intervention based on the patient’s actual needs. The control architecture includes three operating modes: Active Guidance (AG), Assist-As-Needed (AAN), and Restrictive Protection (RP), with automatic transitions based on tracking error. The force field consists of three components: a normal force that brings the patient back toward the center of the virtual tunnel, a tangential force that provides guidance along the trajectory, and a viscous force that ensures stability and smoothness of movement. The numerical simulation was entirely developed in Python environment, using the TRAC-IK library for inverse kinematics and Pinocchio for manipulator dynamics computation. The adopted approach allows integration of the equations of motion in joint space, ensuring physical consistency and numerical accuracy. The results demonstrate the correct functioning of the control system, with a balanced distribution between AG and AAN modes, actuator torques well within safety limits, and dynamic behavior consistent with theoretical expectations.
La riabilitazione robotica degli arti superiori rappresenta un campo in rapida evoluzione, dove l’interazione uomo-robot gioca un ruolo fondamentale nel recupero motorio dei pazienti. Il presente lavoro di tesi si propone di sviluppare e validare un sistema di controllo in impedenza per il robot riabilitativo TIAGo, implementando una strategia di assistenza graduata basata su force field. Il sistema proposto si basa sul paradigma Assist-As-Needed, dove il robot mantiene un comportamento prevalentemente passivo, lasciando al paziente il ruolo di motore primario del movimento. L’assistenza viene fornita in modo graduato attraverso un campo di forze virtuali costruito attorno alla traiettoria di riferimento, modulando l’intervento in base alle effettive necessità del paziente. L’architettura di controllo prevede tre modalità operative: Active Guidance (AG), Assist-As-Needed (AAN) e Restrictive Protection (RP), con transizioni automatiche basate sull’errore di tracking. Il force field è composto da tre componenti: una forza normale che riporta il paziente verso il centro del tunnel virtuale, una forza tangenziale che fornisce guida lungo la traiettoria, e una forza viscosa che garantisce stabilità e fluidità del movimento. La simulazione numerica è stata realizzata interamente in ambiente Python, utilizzando la libreria TRAC-IK per la cinematica inversa e Pinocchio per il calcolo della dinamica del manipolatore. L’approccio adottato consente di integrare le equazioni del moto nello spazio dei giunti, garantendo coerenza fisica e accuratezza numerica. I risultati ottenuti dimostrano il corretto funzionamento del sistema di controllo, con una distribuzione bilanciata tra le modalità AG e AAN, coppie agli attuatori ampiamente entro i limiti di sicurezza, e un comportamento dinamico coerente con le aspettative teoriche.
CONTROLLO IN IMPEDENZA BASATO SU FORCE FIELD CON MODALITÀ DI TRAINING ADATTIVE PER ROBOT RIABILITATIVO TIAGO
GAGLIARDINI, MARCO
2025/2026
Abstract
Robotic rehabilitation of upper limbs represents a rapidly evolving field, where human-robot interaction plays a fundamental role in patients’ motor recovery. This thesis aims to develop and validate an impedance control system for the TIAGo rehabilitation robot, implementing a graduated assistance strategy based on force fields. The proposed system is based on the Assist-As-Needed paradigm, where the robot maintains a predominantly passive behavior, leaving the patient as the primary driver of movement. Assistance is provided gradually through a virtual force field constructed around the reference trajectory, modulating intervention based on the patient’s actual needs. The control architecture includes three operating modes: Active Guidance (AG), Assist-As-Needed (AAN), and Restrictive Protection (RP), with automatic transitions based on tracking error. The force field consists of three components: a normal force that brings the patient back toward the center of the virtual tunnel, a tangential force that provides guidance along the trajectory, and a viscous force that ensures stability and smoothness of movement. The numerical simulation was entirely developed in Python environment, using the TRAC-IK library for inverse kinematics and Pinocchio for manipulator dynamics computation. The adopted approach allows integration of the equations of motion in joint space, ensuring physical consistency and numerical accuracy. The results demonstrate the correct functioning of the control system, with a balanced distribution between AG and AAN modes, actuator torques well within safety limits, and dynamic behavior consistent with theoretical expectations.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi_Gagliardini_ControlloImpedenzaTiago_PDFA.pdf
accesso aperto
Dimensione
3.71 MB
Formato
Adobe PDF
|
3.71 MB | Adobe PDF | Visualizza/Apri |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.12075/28043