This thesis work focuses on the modeling and robust control of nano-quadcopters, with specific reference to the Crazyflie 2.1+ open-source platform, operating in environments characterized by exogenous disturbances and parametric uncertainties. To address the inherently non-linear, underactuated, and coupled nature of the system, a comprehensive dynamic model based on the Newton-Euler formalism was developed and validated, and subsequently linearized around the steady-state flight condition (hovering). This approach enabled the decoupling of the dynamics into four independent subsystems. The propulsion parameters, including the thrust coefficient and the PWM-to-angular velocity calibration curve, were identified via least-squares regression based on experimental data to ensure high actuation fidelity. The proposed control architecture is based on a hierarchical cascade structure operating in an on-board/off-board logic: the outer position and altitude loops generate the kinematic and attitude references, which the inner loop then converts into algebraic actuation commands for the motors. Particular attention was paid to introducing a robust integral action in the planar positioner and implementing an augmented state with 18 variables to eliminate steady-state errors. The effectiveness of the strategy was verified through extensive simulation campaigns in the MATLAB environment. The main results highlight an excellent step-response agility and a high capability to reject both impulsive aerodynamic disturbances (wind gusts) and persistent ones, demonstrating the achievement of global asymptotic stability even in critical scenarios involving sudden disturbance inversion (Wind-Switch). Furthermore, the system guaranteed accurate tracking of three-dimensional circular trajectories subjected to stochastic turbulence. The primary contribution of this work lies in the methodological and numerical alignment between motor physics and software constraints, providing a high-fidelity simulator ready for hardware validation. The results obtained lay a solid foundation for future developments, oriented toward implementation on real aircraft, the integration of state estimators, and the adoption of non-linear predictive control techniques.
Il presente lavoro di tesi si inserisce nel contesto della modellazione e del controllo robusto di nano-quadricotteri, con particolare riferimento alla piattaforma open-source Crazyflie 2.1+, operante in ambienti caratterizzati da disturbi esogeni ed incertezze parametriche. Per far fronte alla natura intrinsecamente non lineare, sottoattuata e accoppiata del sistema, è stato sviluppato e validato un modello dinamico completo basato sul formalismo di Newton-Eulero, successivamente linearizzato attorno alla condizione di volo stazionario (hovering). Tale approccio ha consentito il disaccoppiamento della dinamica in quattro sottosistemi indipendenti. I parametri propulsivi, tra cui il coefficiente di spinta e la curva di calibrazione PWM-velocità angolare, sono stati identificati tramite regressione ai minimi quadrati basata sui dati sperimentali al fine di garantire un'elevata fedeltà attuativa. L'architettura di controllo proposta si basa su una struttura gerarchica a cascata operante in logica on-board/off-board: i loop esterni di posizione e quota generano i riferimenti cinematici e d'assetto che il ciclo interno converte in comandi algebrici di attuazione per i motori. Particolare attenzione è stata rivolta all'introduzione di un'azione integrale robusta nel posizionatore planare e all'implementazione di uno stato aumentato a 18 variabili per annullare gli errori a regime. La bontà della strategia è stata verificata attraverso estensive campagne di simulazione in ambiente MATLAB. I risultati principali evidenziano un'ottima prontezza della risposta al gradino, un'elevata capacità di rigetto di disturbi aerodinamici sia impulsivi (raffica di vento) sia persistenti, dimostrando il raggiungimento della stabilità asintotica globale anche in scenari critici di inversione repentina del disturbo (Wind-Switch). Inoltre, il sistema ha garantito un inseguimento accurato di traiettorie circolari tridimensionali soggette a turbolenza stocastica. Il contributo principale del lavoro risiede nell'allineamento metodologico e numerico tra la fisica dei motori e i vincoli software, fornendo un simulatore ad alta fedeltà pronto per la validazione hardware. I risultati ottenuti pongono solide basi per futuri sviluppi, orientati all'implementazione su velivoli reali, all'integrazione di stimatori dello stato e all'adozione di tecniche di controllo predittivo non lineare.
MODELLAZIONE E CONTROLLO ROBUSTO DELL'HOVERING DI UN NANO-QUADRICOTTERO IN PRESENZA DI DISTURBI ESTERNI
CASUCCI, NUNZIO
2025/2026
Abstract
This thesis work focuses on the modeling and robust control of nano-quadcopters, with specific reference to the Crazyflie 2.1+ open-source platform, operating in environments characterized by exogenous disturbances and parametric uncertainties. To address the inherently non-linear, underactuated, and coupled nature of the system, a comprehensive dynamic model based on the Newton-Euler formalism was developed and validated, and subsequently linearized around the steady-state flight condition (hovering). This approach enabled the decoupling of the dynamics into four independent subsystems. The propulsion parameters, including the thrust coefficient and the PWM-to-angular velocity calibration curve, were identified via least-squares regression based on experimental data to ensure high actuation fidelity. The proposed control architecture is based on a hierarchical cascade structure operating in an on-board/off-board logic: the outer position and altitude loops generate the kinematic and attitude references, which the inner loop then converts into algebraic actuation commands for the motors. Particular attention was paid to introducing a robust integral action in the planar positioner and implementing an augmented state with 18 variables to eliminate steady-state errors. The effectiveness of the strategy was verified through extensive simulation campaigns in the MATLAB environment. The main results highlight an excellent step-response agility and a high capability to reject both impulsive aerodynamic disturbances (wind gusts) and persistent ones, demonstrating the achievement of global asymptotic stability even in critical scenarios involving sudden disturbance inversion (Wind-Switch). Furthermore, the system guaranteed accurate tracking of three-dimensional circular trajectories subjected to stochastic turbulence. The primary contribution of this work lies in the methodological and numerical alignment between motor physics and software constraints, providing a high-fidelity simulator ready for hardware validation. The results obtained lay a solid foundation for future developments, oriented toward implementation on real aircraft, the integration of state estimators, and the adoption of non-linear predictive control techniques.| File | Dimensione | Formato | |
|---|---|---|---|
|
Tesi_Casucci Nunzio_Definitiva.pdf
non disponibili
Dimensione
2 MB
Formato
Adobe PDF
|
2 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14253/68107