This thesis presents the implementation of an Augmented P-MRAC controller on a vector-thrust quadrotor UAV and the analysis of its performance under various fault and disturbance conditions. First, the general Augmented P-MRAC framework adopted in this thesis is introduced, described, and analyzed for a generic uncertain dynamical system. Particular attention is devoted to the structural properties of the system matrices (e.g. overactuation), which are exploited to improve the adaptive laws. Subsequently, the model of a vector-thrust quadrotor UAV is introduced, and two alternative Augmented P-MRAC controller architectures for the uncertain attitude and position dynamics are presented. The first architecture is based on the use of a single augmented P-MRAC controller, in which the uncertain attitude and position dynamics are combined into a unified model. In this case, an adaptive control allocation algorithm is employed, exploiting the online estimation of the efficiency of each individual actuator, to optimize the control allocation process. The second architecture is based on the application of the Augmented P-MRAC controller to each subsystem separately, thereby ensuring higher modularity, reduced computational demand, and a decrease in the number of adaptive parameters needed to be estimated. This reduction is achieved through the projection of the control effectiveness matrix onto the non-null subspace in the case of an overactuated subsystem. Subsequently, an adaptive control allocation algorithm is implemented for real-time applications, combining the two projected estimates of the control effectiveness matrix to reconstruct the full control effectiveness matrix. The implemented Augmented P-MRAC approaches are then validated and compared in a numerical simulation environment (Simulink) under different actuator degradation and disturbance scenarios. Finally, the Augmented P-MRAC controller is implemented on a physical quadrotor platform, and real-world experimental tests are conducted to assess its practical effectiveness.
Questa tesi presenta l’implementazione di un controllore augmented P-MRAC su un UAV quadrirotore a spinta vettoriale e l’analisi delle sue prestazioni in presenza di diverse condizioni di guasto e disturbo. Inizialmente, il framework generale dell’augmented P-MRAC adottato, viene introdotto, descritto e analizzato con riferimento a un sistema dinamico generico incerto. Particolare attenzione è rivolta alle proprietà strutturali delle matrici del sistema (e.g. la sovra-attuazione), che vengono sfruttate per ottimizzare le leggi di adattamento. Successivamente, viene introdotto il modello di un UAV quadrirotore e vengono presentate due architetture alternative di un augmented P-MRAC per le dinamiche incerte di assetto e posizione. La prima architettura si basa sull’utilizzo di un unico augmented P-MRAC, nel quale le dinamiche incerte di assetto e posizione sono combinate in un modello unificato. In questo caso, viene impiegato un algoritmo di allocazione del controllo adattativo, che sfrutta la stima online dell’efficienza di ciascun attuatore al fine di ottimizzare il processo di allocazione del controllo. La seconda architettura si basa sull’applicazione dell’augmented P-MRAC a ciascun sottosistema separatamente, garantendo una maggiore modularità, un minore carico computazionale e una riduzione del numero di parametri adattativi da stimare. Tale riduzione è ottenuta attraverso la proiezione della matrice di efficacia del controllo sul sottospazio non nullo nel caso di un sottosistema sovraattuato. Successivamente, viene implementato un algoritmo di allocazione del controllo adattativo per applicazioni in tempo reale, che combina le due stime della proiezione della matrice di efficacia del controllo al fine di ricostruire la matrice completa di efficacia del controllo. I due diversi approcci P-MRAC implementati, vengono quindi validati e confrontati in un ambiente di simulazione numerica (Simulink) sotto diversi scenari di riduzione dell’efficacia degli attuatori e di disturbo. Infine, l’augmented P-MRAC viene implementato su una piattaforma quadrirotore reale e vengono condotti test sperimentali in condizioni operative reali per valutarne l’efficacia pratica.
Predictor-based adaptive control and allocation for multirotor UAVs
BORSA, NICOLÓ
2024/2025
Abstract
This thesis presents the implementation of an Augmented P-MRAC controller on a vector-thrust quadrotor UAV and the analysis of its performance under various fault and disturbance conditions. First, the general Augmented P-MRAC framework adopted in this thesis is introduced, described, and analyzed for a generic uncertain dynamical system. Particular attention is devoted to the structural properties of the system matrices (e.g. overactuation), which are exploited to improve the adaptive laws. Subsequently, the model of a vector-thrust quadrotor UAV is introduced, and two alternative Augmented P-MRAC controller architectures for the uncertain attitude and position dynamics are presented. The first architecture is based on the use of a single augmented P-MRAC controller, in which the uncertain attitude and position dynamics are combined into a unified model. In this case, an adaptive control allocation algorithm is employed, exploiting the online estimation of the efficiency of each individual actuator, to optimize the control allocation process. The second architecture is based on the application of the Augmented P-MRAC controller to each subsystem separately, thereby ensuring higher modularity, reduced computational demand, and a decrease in the number of adaptive parameters needed to be estimated. This reduction is achieved through the projection of the control effectiveness matrix onto the non-null subspace in the case of an overactuated subsystem. Subsequently, an adaptive control allocation algorithm is implemented for real-time applications, combining the two projected estimates of the control effectiveness matrix to reconstruct the full control effectiveness matrix. The implemented Augmented P-MRAC approaches are then validated and compared in a numerical simulation environment (Simulink) under different actuator degradation and disturbance scenarios. Finally, the Augmented P-MRAC controller is implemented on a physical quadrotor platform, and real-world experimental tests are conducted to assess its practical effectiveness.| File | Dimensione | Formato | |
|---|---|---|---|
|
2026_03_Borsa_Tesi.pdf
accessibile in internet per tutti
Descrizione: Testo Tesi
Dimensione
15.12 MB
Formato
Adobe PDF
|
15.12 MB | Adobe PDF | Visualizza/Apri |
|
2026_03_Borsa_Executive_Summary.pdf
accessibile in internet per tutti
Descrizione: Executive Summary della Tesi
Dimensione
1.67 MB
Formato
Adobe PDF
|
1.67 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/251645