In this dissertation we present advances in developing an autonomous air vehicle capable of navigating through unknown, indoor environments. The problem imposes stringent limits on the computational power available onboard the vehicle, but the environment necessitates using 3D sensors such as stereo or RGB-D cameras whose data requires significant processing. We address the problem by proposing and developing key elements of a relative navigation scheme that moves as many processing tasks as possible out of the time-critical functions needed to maintain flight. We present in Chapter 2 analysis and results for an improved multirotor helicopter state estimator. The filter generates more accurate estimates by using an improved dynamic model for the vehicle and by properly accounting for the correlations that exist in the uncertainty during state propagation. As a result, the filter can rely more heavily on frequent and easy to process measurements from gyroscopes and accelerometers, making it more robust to error in the processing intensive information received from the exteroceptive sensors. In Chapter 3 we present BERT, a novel approach to map optimization. The goal of map optimization is to produce an accurate global map of the environment by refining the relative pose transformation estimates generated by the real-time navigation system. We develop BERT to jointly optimize the global poses and relative transformations. BERT exploits properties of independence and conditional independence to allow new information to efficiently flow through the network of transformations. We show that BERT achieves the same final solution as a leading iterative optimization algorithm. However, BERT delivers noticeably better intermediate results for the relative transformation estimates. The improved intermediate results, along with more readily available covariance estimates, make BERT especially applicable to our problem where computational resources are limited. We conclude in Chapter 4 with analysis and results that extend BERT beyond the simple example of Chapter 3. We identify important structure in the network of transformations and address challenges arising in more general map optimization problems. We demonstrate results from several variations of the algorithm and conclude the dissertation with a roadmap for future work.

Optimal control for multicopters is difficult in part due to the low processing power available, and the instability inherent to multicopters. Deep imitation learning is a method for approximating an expert control policy with a neural network, and has the potential of improving control for multicopters. We investigate the performance and reliability of deep imitation learning with trajectory optimization as the expert policy by first defining a dynamics model for multicopters and applying a trajectory optimization algorithm to it. Our investigation shows that network architecture plays an important role in the characteristics of both the learning process and the resulting control policy, and that in particular trajectory optimization can be leveraged to improve convergence times for imitation learning. Finally, we identify some limitations and future areas of study and development for the technology. / Optimal kontroll för multikoptrar är ett svårt problem delvis på grund av den vanligtvis låga processorkraft som styrdatorn har, samt att multikoptrar är synnerligen instabila system. Djup imitationsinlärning är en metod där en beräkningstung expert approximeras med ett neuralt nätverk, och gör det därigenom möjligt att köra dessa tunga experter som realtidskontroll för multikoptrar. I detta arbete undersöks prestandan och pålitligheten hos djup imitationsinlärning med banoptimering som expert genom att först definiera en dynamisk modell för multikoptrar, sedan applicera en välkänd banoptimeringsmetod på denna modell, och till sist approximera denna expert med imitationsinlärning. Vår undersökning visar att nätverksarkitekturen spelar en avgörande roll för karakteristiken hos både inlärningsprocessens konvergenstid, såväl som den resulterande kontrollpolicyn, och att särskilt banoptimering kan nyttjas för att förbättra konvergenstiden hos imitationsinlärningen. Till sist påpekar vi några begränsningar hos metoden och identifierar särskilt intressanta områden för framtida studier.

This thesis presents the initial components of an integrated guidance, navigation, and control system for vertical take-off and landing (VTOL) autonomous unmanned aerial vehicles (UAVs) such that they may map complex environments that may be hostile. The first part of this thesis presents an autonomous guidance system. For goal selection, the map is partitioned around the presence of obstacles and whether that area has been explored. To perform this partitioning, the Octree algorithm is implemented. In this thesis, we test this algorithm to find a parameter set that optimizes this algorithm. Having selected goal points, we perform a comparison of the LPA* and A* path planning algorithms with a custom heuristic that enables reckless or tactical maneuvers as the UAV maps the environment. For trajectory planning, the fMPC algorithm is applied to the feedback-linearized equations of motion of a quadcopter. For collision avoidance, standalone versions of 4 different constraint generation algorithms are evaluated to compare their resulting computation times, accuracy, and computed volume on a voxel map that simulates a 2-story house along with fixed paths that vary in length at fixed intervals as basis of tests. The second part of this thesis presents the theory of Model Reference Adaptive Control(MRAC) along with augmentation for output signal tracking and switched-dynamic systems. We then detail the development of longitudinal and lateral controllers a Quad-Rotor Tailsitter(QRBP) style UAV. In order to successfully implement the proposed controller on the QRBP, significant effort was placed upon physical design and testing apparatus. / Master of Science / For an autonomously operated, Unmanned Aerial Vehicle (UAV), to operate, it requires a guidance system to determine where and how to go, and a control system to effectively actuate the guidance system's commands. In this thesis, we detail the characterization and optimization of the algorithms comprising the guidance system. We then delve into the theory of MRAC and apply it toward a control system for a QRBP. We then detail additional tools developed to support the testing of the QRBP.

Le travail présenté dans ce mémoire concerne le développement théorique et la validation expérimentale d'algorithmes de fusion de données originaux pour mini-drones, dépassant les limitations des estimateurs communément utilisés (e.g. le Filtre de Kalman Etendu ou les filtres particulaires) et pouvant être implémentés aisément et efficacement dans une avionique bas-coûts. Nous proposons tout d'abord des observateurs invariants "génériques'', préservant les symétries naturelles du système physique. Ces observateurs fusionnent les mesures de capteurs bon marché usuels (tels que les capteurs inertiels, magnétomètres, GPS ou baromètres) afin d'estimer avec précision l'état de l'appareil (angles d'attitude et de cap, vitesse et position) sans modèle dynamique de l'engin. Ils possèdent un large domaine de convergence; ils sont également faciles à régler et très économiques en temps de calcul. Puis nous développons des observateurs "spécifiques'', adaptés au type de véhicule aérien considéré, en l'occurence un mini-quadrotor. Basés sur un modèle physique tenant compte de la traînée de rotor, les observateurs présentés estiment la vitesse du quadrotor à partir de mesures uniquement inertielles, menant à un contrôle en vitesse de l'appareil. Cette approche est validée par des vols stabilisés autonomes. Enfin, nous présentons en détail l'intégration du système avionique bas-coûts utilisé, composé de capteurs "bruts'' et d'un microcontrôleur sur lequel sont implémentés les observateurs précédents. Nous validons ces algorithmes en comparant leurs estimations avec celles fournies par un produit commercial coûteux, mettant ainsi en évidence leur excellent rapport "qualité/prix''.

La vision représente un des plus importants signaux en robotique. Une caméra monoculaire peut fournir de riches informations visuelles à une fréquence raisonnable pouvant être utilisées pour la commande, l’estimation d’état ou la navigation dans des environnements inconnus par exemple. Il est cependant nécessaire de respecter des contraintes visuelles spécifiques telles que la visibilité de mesures images et les occultations durant le mouvement afin de garder certaines cibles visuelles dans le champ de vision. Les quadrirotors sont dotés de capacités de mouvement très réactives du fait de leur structure compacte et de la configuration des moteurs. De plus, la vision par une caméra embarquée (fixe) va subir des rotations dues au sous-actionnement du système. Dans cette thèsenous voulons bénéficier de l’agilité du quadrirotor pour réaliser plusieurs tâches de navigation basées vision. Nous supposons que l’estimation d’état repose uniquement sur la fusion capteurs d’une centrale inertielle (IMU) et d’une caméra monoculaire qui fournit des estimations de pose précises. Les contraintes visuelles sont donc critiques et difficiles dans un tel contexte. Dans cette thèse nous exploitons l’optimisation numérique pour générer des trajectoires faisables satisfaisant un certain nombre de contraintes d’état, d’entrées et visuelles non linéaires. A l’aide la platitude différentielle et de la paramétrisation par des B-splines nous proposons une stratégie de replanification performante inspirée de la commande prédictive pour générer des trajectoires lisses et agiles. Enfin, nous présentons un algorithme de planification en temps minimum qui supporte des pertes de visibilité intermittentes afin de naviguer dans des environnements encombrés plus vastes. Cette contribution porte l’incertitude de l’estimation d’état au niveau de la planification pour produire des trajectoires robustes et sûres. Les développements théoriques discutés dans cette thèse sont corroborés par des simulations et expériences en utilisant un quadrirotor. Les résultats reportés montrent l’efficacité des techniques proposées. / Vision constitutes one of the most important cues in robotics. A single monocular camera can provide rich visual information at a reasonable rate that can be used as a feedback for control, state estimation of mobile robots or safe navigation in unknown environments for instance. However, it is necessary to satisfy particular visual constraints on the image such as visibility and occlusion constraints during motion to keep some visual targets visible. Quadrotors are endowed with very reactive motion capabilities due to their compact structure and motor configuration. Moreover, vision from a (fixed) on-board camera will suffer from rotation motions due to the system underactuation. In this thesis, we want to benefit from the system aggressiveness to perform several vision-based navigation tasks. We assume state estimation relies solely on sensor fusion of an onboard inertial measurement unit (IMU) and a monocular camera that provides reliable pose estimates. Therefore, visual constraints are challenging and critical in this context. In this thesis we exploit numerical optimization to design feasible trajectories satisfying several state, input and visual nonlinear constraints. With the help of differential flatness and B-spline parametrization we will propose an efficient replanning strategy inspired form Model Predictive Control to generate smooth and agile trajectories. Finally, we propose a minimum-time planning algorithm that handles intermittent visibility losses in order to navigate in larger cluttered environments. This contribution brings state estimation uncertainty at the planning stage to produce robust and safe trajectories. All the theoretical developments discussed in this thesis are corroborated by simulations and experiments run by using a quadrotor UAV. The reported results show the effectiveness of proposed techniques.

This thesis presents a real-time autonomous guidance and control method for a quadrotor in a GPS-denied environment. The quadrotor autonomously seeks a destination while it avoids obstacles whose shape and position are initially unknown. We implement the obstacle avoidance and destination seeking methods using off-the-shelf sensors, including a vision-sensing camera. The vision-sensing camera detects the positions of points on the surface of obstacles. We use this obstacle position data and a potential-field method to generate velocity commands. We present a backstepping controller that uses the velocity commands to generate the quadrotor's control inputs. In indoor experiments, we demonstrate that the guidance and control methods provide the quadrotor with sufficient autonomy to fly point to point, while avoiding obstacles.

Unmanned Aerial Vehicles are often used for reconnaissance, search and rescue, damage assessment, exploration, and other tasks that are dangerous or prohibitively difficult for humans to perform. Often, these tasks include traversing indoor environments where radio links are unreliable, hindering the use of remote pilot links or ground-based control, and effectively eliminating Global Positioning System (GPS) signals as a potential localization method. As a result, any vehicle capable of indoor flight must be able to stabilize itself and perform all guidance, navigation, and control tasks without dependence on a radio link, which may be available only intermittently. Since the availability of GPS signals in unknown environments is not assured, other sensors must be used to provide position information relative to the environment. This research covers a description of different ranging sensors and methods for incorporating them into the overall guidance, navigation, and control system of a flying vehicle. Various sensors are analyzed to determine their performance characteristics and suitability for indoor navigation, including sonar, infrared range sensors, and a scanning laser rangefinder. Each type of range sensor tested has its own unique characteristics and contributes in a slightly different way to effectively eliminate the dependence on GPS. The use of low-cost range sensors on an inexpensive passively stabilized coaxial helicopter for drift-tolerant indoor navigation is demonstrated through simulation and flight test. In addition, a higher fidelity scanning laser rangefinder is simulated with an Inertial Measurement Unit (IMU) onboard a quadrotor helicopter to enable active stabilization and position control. Two different navigation algorithms that utilize a scanning laser and techniques borrowed from Simultaneous Localization and Mapping (SLAM) are evaluated for use with an IMU-stabilized flying vehicle. Simulation and experimental results are presented for each of the navigation systems.

[pt] Quadrirrotores são veículos aéreos que possuem quatro rotores fixos e orientados na direção vertical. Devido à sua simplicidade mecânica frente aos helicópteros tradicionais, os mesmos têm se tornado cada vez mais populares nos meios de pesquisa, militares e, mais recentemente, industriais. Essa topologia de veículo data do início do século XX mas o desenvolvimento em escala só foi possível após a recente evolução e miniaturização dos sistemas eletrônicos embarcados, dos motores elétricos e das baterias. A movimentação desses veículos no espaço é possível graças à sua inclinação em relação ao solo e, para tal, é imprescindível obter e controlar corretamente a atitude do mesmo. As unidades de medidas inerciais (IMU) surgiram como uma solução para esse problema. Através da fusão dos dados obtidos com os sensores presentes nessas centrais (acelerômetros, girômetros e magnetômetro) é possível estimar a atitude do veículo. O presente trabalho apresenta soluções tanto para a estimativa quanto para o controle de atitude de quadrirrotor. Os modelos matemáticos desenvolvidos são validados em simulações numéricas e em testes experimentais. O objetivo é que as soluções propostas apresentem resultados positivos para que possam ser empregadas nos quadrirrotores em escala. / [en] Quadrotors are vehicles that have four fixed rotors in the vertical direction. Due to its mechanical simplicity compared to traditional helicopters, these vehicles have become increasingly popular in the research, military and, more recently, industrial fields. This type of vehicle first appeared in the early twentieth century, but the development of small-scale models was only possible after the recent evolution and miniaturization of embedded electronics, electric motors and batteries. A Quadrotor can fly in any direction by changing its inclination relative to the ground, so it is essential to calculate and properly adjust its attitude. The inertial measurement units (IMU) emerged as one solution to this problem. By merging the IMU sensors data, it is possible to estimate the vehicle s attitude. This dissertation presents solutions for both the estimation and the control of the vehicle s attitude. The developed mathematical models are validated with numerical simulations and experimental tests. The goal is that the presented solutions give enough good results so they can be used in small-scale Quadrotors.

Orientador: Prof. Dr. André Fenili / Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Engenharia Mecânica, Santo André, 2018.

Cette thèse étudie certains problèmes plus importants dans le sens de guidage, navigation et contrôle présentés dans une catégorie particulière de mini véhicules aériens (MVA) : le MVA convertible avec des ailes fixes et disques pendulaires. Cet aéronef est capable de changer sa configuration de vol, du vol stationnaire au vol palier et vice versa, au moyen d’une manœuvre de transition. Motivé par des applications civiles, on étudie théoriquement et expérimentalement les principes de contrôle en fonction de Lyapunov pour les dynamiques présentées dans le MVA convertible. Des résultats de convergence asymptotique sont obtenus sur l’enveloppe de vol complet du véhicule : d’un vol vertical à basse vitesse à un vol vers l’avant à grande vitesse. Cette thèse est divisée en quatre parties principales : l’étude de 1) les aéronefs à voilure fixe ; 2) le quadrirotor (avion équipe de quatre moteurs) ; 3) l’aéronef convertible ; 4) les applications de vision en utilisant l’aéronef convertible. Dans la première partie, un principe de contrôle en fonction de Lyapunov est développé pour diriger un mini véhicule aérien à voilure fixe tout au long d’un chemin d’accès souhaité. En outre, un générateur de chemin d’accès est proposé. Le résultant de la stratégie du contrôle donne une convergence globale du chemin actuel du MVA au chemin d’accès souhaité. Dans la deuxième partie, un contrôle en fonction de Lyapunov à l’aide de la théorie de la perturbation du singulier est proposé et appliqué sur la dynamique du MVA. En effet, dans cette partie on a abordé le problème diagnostic et la détection de pannes fault detection and diagnosis (FDD) pour un quadrirotor. Dans la troisième partie une nouvelle stratégie de contrôle pour effectuer la transition d’un avion convertible entre le mode avion et le mode hélicoptère, et vice versa, est présenté. L’analyse est effectuée pour le modèle longitudinal du PVHAT (Planar Vertical Helicopter-Airplane Transition) aéronef, lequel est un avion ayant disques pendulaires afin de réaliser la manœuvre de transition. L’algorithme de contrôle de boucle fermée qui en résulte, est prouvé être globalement asymptotiquement stable. Finalement, dans la quatrième partie de cette thèse, le problème de l’estimation et suivi d’un chemin à l’aide de vision système embarqué dans l’avion PVHAT est résolu. La stabilité globale exponentielle de la position sous-système ainsi que le contrôleur de commutation est démontrée. Des simulations illustratives et résultats expérimentaux sont obtenus sur plusieurs plateformes expérimentales développées dans cette thèse, pour évaluer l’applicabilité des principes contrôle proposés et mettre en valeur les mérites de l’approche. / This thesis studies some of the most relevant problems in the sense of guidance,navigation and control presented in a particular class of mini aerial vehicles (MAV) : the convertible MAV with fixed wings and tilting rotors. This aircraft is able to change its flight configuration from hover to level flight and vice-versaby means of a transition maneuver. Motivated by civilian applications, we theoretically and experimentally study Lyapunov-based control laws for dynamics presented in the convertible MAV. Results of asymptotic convergence are obtained over the complete flight envelope of the vehicle : from low-speed vertical flight through high-speed forward flight. We have divided this thesis in four main parts : the study of 1) the fixed-wingaircraft; 2) the quadrotor; 3) the convertible aircraft and 4) vision applications by using the convertible aircraft. In a first part, a Lyapunov-based controllaw is developed to steer a fixed wing mini aerial vehicle along a desired path. Furthermore a path generator is proposed. The resulting control strategy yields global convergence of the current path of the MAV to the desired path. In a second part, a Lyapunov-based control using singular perturbation theory is proposed and applied on dynamics of the MAV. Furthermore, in this part we address the problem of fault detection and diagnosis (FDD) for a quad-rotor. In the third part a new control strategy for the transition between airplane and helicopter mode, and vice versa, in convertible planes is presented. The analysis is carried out for the longitudinal model of the PVHAT (Planar VerticalHelicopter-Airplane Transition) aircraft, which is an airplane having tilting rotors in order to achieve the transition maneuver. The resulting closed loop control algorithm is proved to be globally asymptotically stable. Finally in thefourth part of this thesis the problem of estimation and tracking of a road using avision embedded system in the PVHAT aircraft is solved. The global exponential stability of the position subsystem together with the switching controller is demonstrated. Illustrative simulations and experimental results obtained on several experimental platforms developed in this thesis, assess the implementability of the proposed control laws and highlight the merits of the approach.