The role of vision as an additional sensing mechanism has received a lot of attention in recent years in the context of autonomous flight applications. Modern Unmanned Aerial Vehicles (UAVs) are equipped with vision sensors because of their light-weight, low-cost characteristics and also their ability to provide a rich variety of information of the environment in which the UAVs are navigating in. The problem of vision based autonomous flight is very difficult and challenging since it requires bringing together concepts from image processing and computer vision, target tracking and state estimation, and flight guidance and control. This thesis focuses on the adaptive state estimation, guidance and control problems involved in vision-based formation flight. Specifically, the thesis presents a composite adaptation approach to the partial state estimation of a class of nonlinear systems with unmodeled dynamics. In this approach, a linear time-varying Kalman filter is the nominal state estimator which is augmented by the output of an adaptive neural network (NN) that is trained with two error signals. The benefit of the proposed approach is in its faster and more accurate adaptation to the modeling errors over a conventional approach. The thesis also presents two approaches to the design of adaptive guidance and control (G&C) laws for line-of-sight formation flight. In the first approach, the guidance and autopilot systems are designed separately and then combined together by assuming time-scale separation. The second approach is based on integrating the guidance and autopilot design process. The developed G&C laws using both approaches are adaptive to unmodeled leader aircraft acceleration and to own aircraft aerodynamic uncertainties. The thesis also presents theoretical justification based on Lyapunov-like stability analysis for integrating the adaptive state estimation and adaptive G&C designs. All the developed designs are validated in nonlinear, 6DOF fixed-wing aircraft simulations. Finally, the thesis presents a decentralized coordination strategy for vision-based multiple-aircraft formation control. In this approach, each aircraft in formation regulates range from up to two nearest neighboring aircraft while simultaneously tracking nominal desired trajectories common to all aircraft and avoiding static obstacles.

In this thesis, a new guidance method for fighter aircrafts and a new guidance method for missiles are developed. Also, guidance and control systems of the aircraft and the missile used are designed to simulate the generic engagement scenarios between the missile and the aircraft. Suggested methods have been tested under excessive simulation studies. The aircraft guidance method developed here is a rule based missile evasion method. The main idea to develop this method stems from the maximization of the miss distance for an engagement scenario between a missile and an aircraft. To do this, an optimal control problem with state and input dependent inequality constraints is solved and the solution method is applied on different problems that represent generic scenarios. Then, the solutions of the optimal control problems are used to extract rules. Finally, a method that uses the interpolation of the extracted rules is given to guide the aircraft. The new guidance method developed for missiles is formulated by modifying the classical proportional navigation guidance method using the position estimates. The position estimation is obtained by utilization of a Kalman based filtering method, called interacting multiple models.

Kinematics based guidance laws like the proportional navigation (PN) and many other linear optimal guidance laws perform well in near-collision course conditions. These have been studied thoroughly in the literature from all aspects, ranging from optimality to capturability, for planar or two dimensional interceptor-target engagements, and to a lesser extent, for three dimensional engagements. But guidance in widely off-collision course conditions like high initial heading errors has been relatively less studied. This is probably due to the inherently high nonlinearity of the problem, which makes it a far more difficult problem to solve. However, with increasing speed and agility of interceptors and targets, solutions of such problems have acquired an increased urgency, as has been reflected in the recent literature. This thesis proposes a guidance law based on differential evolution (DE), a member of the evolutionary algorithms (EA) family. While EAs have been applied extensively to static optimization problems, they have been considered unsuitable for solving dynamic optimization or optimal control problems, due to their computationally intensive nature, and their consequent inability to produce solutions online in real-time, except for systems with very slow dynamics. This thesis proposes an online-implementable optimal control for interceptor guidance, a problem with inherently fast dynamics. The proposed law is applicable to all initial geometries including those that involve high to very high heading errors. While interception by itself is a challenging task in the presence of high heading errors, an additional requirement of optimality is also imposed. The first part of the thesis considers only the 2-D kinematic model with high heading errors. In the second part, a 3-D realistic dynamic model, which includes a time-varying interceptor speed, thrust, drag and mass, besides gravity in the vertical plane of motion, and upper bound on the lateral acceleration, is considered, in addition to high heading errors. It is shown that the same structure of the law that is proposed for the 2-D kinematic model can also be used for the 3-D realistic model, if the rest of the complexities of moving from 2-D space to 3-D space, and from kinematics to dynamics is duly addressed. The guidance law proposed does not require time-to-go, the estimation of which can be a difficult problem in high heading error scenarios in which the closing velocity can be negative. Easy to compute and simple to implement in practice, the proposed law does not need any of the techniques or methods from classical optimal control theory, which are complicated and suffer from several limitations. The empirical pure PN (PPN) law is augmented with a term that is a polynomial function of the heading error. The values of the coefficients of the polynomial are found by using the DE. The computational effort required for this low dimensional polynomial optimization problem is shown to be low enough to enable online implementation in real-time. The performance of the proposed law in nominal and off-nominal conditions is validated through several simulations for the 2-D kinematic model, and the 3-D realistic dynamic model. The results are compared with the PPN, augmented PPN and the all-aspect proportional navigation (AAPN) laws in the literature, as per several criteria like optimality, peak latax required and robustness to off-nominal conditions. A successful online implementation of the proposed law for application in practice is also demonstrated. An obvious limitation of optimization by soft computation methods like differential evolution is that no rigorous proof of either convergence or optimality exists for such methods. A fallback option in the form of a conventional guidance law is included in the scheme in case of failure of convergence, and an indirect proof of optimality is provided in the third and final part of the thesis. The same guidance problem is solved by direct multiple shooting method, and it is shown that the numerical results of the two methods compare favourably. The solution by the shooting method is optimal, but computationally far more intensive and takes a computation time of an order of magnitude that is at least one or two times that of the simulation time of the plant. It also needs a good initial guess solution that lies within the region of convergence, which can be a difficult task by itself. Moreover, the shooting method solution is only open loop, and hence applicable for the given model and initial conditions only. Whereas, the simplicity of the method proposed in the thesis makes the solution or guidance law computable in a fraction of the flight time of the engagement, thereby making it online implementable. Equally important, is the fact that it is closed loop, and hence robust to off-nominal conditions like variations in the plant model and parameters assumed in its design.

No description available.

La première année de thèse a permis de mettre en avant deux aspects concernant la problématique de gestion de l’énergie, à savoir le contrôle court terme et le contrôle long terme de l’énergie respectivement. La première problématique a été étudiée pendant la deuxième année de thèse et a débouché sur la proposition d’une architecture de contrôle multi-actionneurs utilisant les moteurs et les aérofreins dans l’objectif d’augmenter l’autorité de contrôle de l’énergie de l’avion. La seconde problématique a été étudiée pendant la troisième année et a débuté par une étude préliminaire reposant sur le calcul d’une séquence optimale de commandes des becs/volets et train d’atterrissage permettant d’amener l’avion à un certain niveau d’énergie en approche tout en minimisant l’utilisation des moteurs et des aérofreins. Par la suite, l’étude a été étendue afin de prendre en compte la régulation des moteurs, l’utilisation des aérofreins et la modification de la trajectoire verticale. Finalement,une solution basée sur un calcul d’optimisation a été développée puis intégrée au sein d’un simulateur de bureau temps-réel, testée avec une interface homme machine adéquat et pour finir présentée à des pilotes d’essais pour validation. / The first year of thesis allowed to foreground two aspects of the energy management issue,namely the short term control and the long term control, respectively. The first issue was studied during the second year and ended with the proposition of a solution mixing both the airbrakes and engines. The second and last issue started with a preliminary study which consisted of computing an optimal slat/flap command sequence bringing the aircraft to theright energy level in approach while minimizing the use of engines and airbrakes. Then, this study was extended in order to take into account the regulation of engines and airbrakes aswell as vertical trajectory modification. Finally, this optimization-based solution has been integrated within an accurate real-time desktop simulator, tested with a human-machine interface, and then presented to flight test pilots for validation.

<p>Autonomous guidance and control techniques for low-thrust spacecraft under multi-body dynamics via reinforcement learning</p>

The Cal Poly CubeSat Laboratory (CPCL) is currently facing unprecedented engineering challenges—both technically and programmatically—due to the increasing cost and complexity of CubeSat flight missions. In responding to recent RFPs, the CPCL has been forced to find commercially available solutions to entire mission critical spacecraft subsystems such as propulsion and attitude determination & control, because currently no in-house options exist for consideration. The commercially available solutions for these subsystems are often extremely expensive and sometimes provide excessively good performance with respect to mission requirements. Furthermore, use of entire commercial subsystems detracts from the hands-on learning objectives of the CPCL by removing engineering responsibility from students. Therefore, if these particular subsystems can be designed, tested, and integrated in-house at Cal Poly, the result would be twofold: 1) the space of missions supportable by the CPCL under tight budget constraints will grow, and 2) students will be provided with unique, hands-on guidance, navigation, and control learning opportunities. In this thesis, the CPCL’s attitude determination and control system design and analysis toolkit is significantly improved to support in-house ADCS development. The toolkit—including the improvements presented in this work—is then used to complete the existing, partially complete CPCL ADCS design. To fill in missing gaps, particular emphasis is placed on guidance and control algorithm design and selection of attitude actuators. Simulation results show that the completed design is competitive for use in a large class of small satellite missions for which pointing accuracy requirements are on the order of a few degrees.

Les travaux présentés dans ce mémoire contribuent à la définition de méthodes pour la détection et le diagnostic de défauts affectant les systèmes aéronautiques. Un système représentatif sert de support d'étude, constitué du modèle non linéaire à six degrés de liberté d'un missile intercepteur, de ses capteurs et actionneurs ainsi que d'une boucle de guidage-pilotage. La première partie est consacrée au développement de deux méthodes de diagnostic exploitant l'information de commande en boucle fermée et les caractéristiques des modèles aéronautiques. La première méthode utilise les objectifs de commande induits par les lois de guidage-pilotage pour générer des résidus indiquant la présence de défauts. Ceci permet la détection des défauts sur les actionneurs et les capteurs, ainsi que leur localisation pour ces derniers. La deuxième méthode exploite la mesure de dérivées des variables d'état (via une centrale inertielle) pour estimer la valeur de la commande réalisée par les actionneurs, sans intégration du modèle non linéaire du système. Le diagnostic est alors effectué en comparant cette estimée avec la valeur désirée, ce qui permet la détection, la localisation et l'identification de défauts multiples sur les actionneurs.La seconde partie propose une méthodologie de réglage automatique des paramètres internes (les hyperparamètres) de méthodes de diagnostic. Ceci permet une comparaison plus objective entre les méthodes en évaluant la meilleure performance de chacune. Le réglage est vu comme un problème d'optimisation globale, la fonction à optimiser étant calculée via la simulation numérique (potentiellement coûteuse) de cas test. La méthodologie proposée est fondée sur un métamodèle de krigeage et une procédure itérative d’optimisation bayésienne, qui permettent d’aborder ce problème à faible coût de calcul. Un nouvel algorithme est proposé afin d'optimiser les hyperparamètres d'une façon robuste vis à vis de la variabilité des cas test pertinents.Mots clés : détection et diagnostic de défauts, guidage-pilotage, krigeage, minimax continu, optimisation globale, redondance analytique, réglage automatique, systèmes aéronautiques. / This manuscript reports contributions to the development of methods for fault detection and diagnosis applied to aeronautical systems. A representative system is considered, composed of the six-degree-of-freedom nonlinear model of a surface-to-air missile, its sensors, actuators and the associated GNC scheme. The first part is devoted to the development of two fault diagnosis approaches that take advantage of closed-loop control information, along with the characteristics of aeronautical models. The first method uses control objectives resulting from guidance laws to generate residuals indicative of the presence of faults. This enables the detection of both actuator and sensor faults, and the isolation of sensor faults. The second method exploits the measurement of derivatives of state variables (as provided by an IMU) to estimate the control input as achieved by actuators, without the need to integrate the nonlinear model. Detection, isolation and identification of actuator faults can then be performed by comparing this estimate with the desired control input.The second part presents a new automatic-tuning methodology for the internal parameters (the hyperparameters) of fault diagnosis methods. This allows a fair comparison between methods by evaluating their best performance. Tuning is formalised as the global optimization of a black-box function that is obtained through the (costly) numerical simulation of a set of test cases. The methodology proposed here is based on Kriging and Bayesian optimization, which make it possible to tackle this problem at a very reduced computational cost. A new algorithm is developed to address the optimization of hyperparameters in a way that is robust to the variability of the test cases of interest.

In this thesis the problem of autonomous landing of an unmanned aerial vehicle named AE-2 is addressed. The guidance and control technique is developed and demonstrated through numerical simulation results. The complete work includes Mathematical modeling, Control design, Guidance and State estimation for AE-2, which is a fixed wing vehicle with 2m wing span and 6kg weight. The aerodynamic data for AE-2 is available from static wind tunnel tests. Functional fit is done on the wind tunnel data with least squares method to find static aerodynamic coefficients. The aerodynamic forces and moment coefficients are highly nonlinear some of them are partitioned in two zones based on the angle of attack. The dynamic derivatives are found with Athena Vortex Lattice software. For the validation of vortex lattice method the static derivatives obtained by the wind tunnel tests and vortex lattice method, are compared before finding dynamic derivatives. The dynamics of the servo actuators for the aerodynamic control surfaces is incorporated in the simulation. The nonlinear dynamic inversion technique has been used for the guidance and control design. The control is structured in two loops, outer and inner loop. The goal of outer loop is to track the guidance commands of altitude, roll angle and yaw angle by converting them into body rate commands through dynamic inversion. The inner loop than tracks these commanded roll rate, pitch rate and yaw rate by finding the required deflection of control surfaces. The forward velocity of the vehicle is controlled by varying the throttle. A controller for actuator is also designed to reduce the lag. The guidance for landing consists of three phases approach, glideslope and flare. During approach the vehicle is aligned with the runway and guided to a specified height from where the glideslope can begin. The glideslope is straight line path specified by a flight path angle which is restricted between 3 to 4 degree. At the end of glideslope which is marked by flare altitude the flare maneuver begins which is an exponential curve. The problem of transition between the glideslope and flare has addressed by ensuring continuity and smoothness at transition. The exponential curve of flare is designed to end below the ground so that it intersects the ground at a prespecified point. The sink rate at touchdown is also controlled along with the location of touchdown point. The state estimation has been done with Extended Kalman Filter in continuous discrete formulation. The external disturbances like wind shear and wind gust are accounted by appending them in state variables. Further the control design with guidance is tested from various initial conditions, in presence of wind disturbances. The designed filter has also been tested for parameter uncertainty.

Die Arbeit befasst sich mit der Entwicklung automatischer Lenkungen, die die von Schienenfahrzeugen bekannte Spurtreue auf n-gliedrige, mehrachsgelenkte Straßenfahrzeuge übertragen. Spurtreu bedeutet folglich, dass die Lenkachsmittelpunkte keinen seitlichen Versatz zueinander aufweisen. Dafür wird ein modellbasiertes automatisches Lenkverfahren systematisch konzipiert, entworfen und erprobt, das sowohl eine vollautomatische Spurführung als auch eine halbautomatische Nachführung erlaubt. Die modellbasierten automatischen Lenkungen unterliegen keinen praktisch relevanten Einschränkungen. Das wird durch die Verwendung einer Steuerungsstruktur mit zwei Freiheitsgraden erreicht, die aus einer modellbasierten Vorsteuerung und einem Rückführregler besteht. In der Vorsteuerung werden die Lenkwinkel aller Achsen berechnet, mit denen der Sollweg theoretisch spurtreu befahren wird. Durch den Einsatz eines speziell angepassten, modularen Mehrkörpermodells gelingt diese Berechnung allgemein für eine Klasse n-gliedriger Fahrzeuge. Zum Ausgleich von nicht vermeidbaren Modellunbestimmtheiten und nicht gemessenen Störungen werden ein nichtlinearer Mehrgrößenregler sowie achs-individuelle lineare Eingrößenregler entworfen und miteinander verglichen. Simulationen und Fahrversuche zeigen, dass das entwickelte Verfahren in einem weiten Geschwindigkeitsbereich robust gegenüber typischen Einflussgrößen wie Fahrbahn- und Beladungszustand ist.