In this dissertation, a novel and generic solution of trajectory generation is developed and evaluated for ground and aerial vehicles in a dynamic environment. By explicitly considering a kinematic model of the ground vehicles, the family of feasible trajectories and their corresponding steering controls are derived in a closed form and are expressed in terms of one adjustable parameter for the purpose of collision avoidance. A collision-avoidance condition is developed for the dynamically changing environment, which consists of a time criterion and a geometrical criterion. By imposing this condition, one can determine a family of collision-free paths in a closed form. Then, optimization problems with respect to different performance indices are setup to obtain optimal solutions from the feasible trajectories. Among these solutions, one with respect to the near-shortest distance and another with respect to the near-minimal control energy are analytical and simple. These properties make them good choices for real-time trajectory planning. Such optimal paths meet all boundary conditions, are twice differentiable, and can be updated in real time once a change in the environment is detected. Then this novel method is extended to 3D space to find a real-time optimal path for aerial vehicles. After that, to reflect the real applications, obstacles are classified to two types: "hard" obstacles that must be avoided, and "soft" obstacles that can be run over/through. Moreover, without losing generality, avoidance criteria are extended to obstacles with any geometric shapes. This dissertation also points out that the emphases of the future work are to consider other constraints such as the bounded velocity and so on. The proposed method is illustrated by computer simulations.

No description available.

No description available.

No description available.

In this dissertation, guidance and control of autonomous unmanned aerial systems (UAS) are explored. Specifically, we investigate model reference adaptive control (MRAC) based systems for tailsitter UAS, and guidance and control of multi-rotor UAS for tactical maneuvering and coverage. Applications, both current and potential, are investigated and gaps in existing technologies are identified. To address the controls problem of a particular class of tailsitter UAS, that is, quadrotor-biplanes, subject to modeling uncertainties, unmodeled payloads, wind gusts, and actuator faults and failures, two approaches are developed. In the first approach, the longitudinal dynamics of a tailsitter UAS are regulated using an MRAC law for prescribed performance and output tracking in a novel control architecture. The MRAC law for prescribed performance and output tracking incorporates a Linear Quadratic Regulator (LQR) baseline controller using integral-feedback interconnections. Constraints on the trajectory tracking error are enforced using barrier Lyapunov functions, and a user-defined rate of convergence of the trajectory tracking error is guaranteed by employing a reference model for the trajectory tracking error's transient dynamics. In this control system, the translational and rotational dynamics are split into an outer loop and an inner loop, respectively, to account for the underactuation of the quadrotor-biplane. In the outer loop, estimates of the aerodynamic forces and MRAC laws are used to stabilize the translational dynamics. Furthermore, the reference pitch angle is deduced such that the vehicle's total thrust never points towards the Earth for safety, and discontinuities inherent to the signed arctangent function commonly used for determining orientations are avoided. In the inner loop, estimates of the aerodynamic moment and an MRAC law are used to stabilize the rotational dynamics. A law for determining the desired total thrust is proposed, which ensures that if the vehicle's orientation is close enough to the desired orientation, then the proper thrust force is applied. A control allocation scheme is presented to ensure that the desired moment of the thrust force is always realized, and constraints on the non-negativity of the thrust force produced by the actuators are satisfied. The proposed control architecture employing MRAC for prescribed performance and output signal tracking is validated in simulation, and the MRAC law for prescribed performance is compared to the classical MRAC law. In the second approach, a unified control architecture based on MRAC is presented which does not separate the longitudinal and lateral-directional dynamics. The translational and rotational dynamics are separated into outer and inner loops, respectively, to address the underactuation of the tailsitter UAS. Since it is expected that the vehicle will undergo large rotations, the tailsitter's orientation is captured using quaternions, which are singularity-free. Furthermore, the windup phenomenon is addressed by employing barrier Lyapunov functions to ensure that the first component of the tracking error quaternion is positive, and thus, the shortest rotation is followed to drive the vehicle's current orientation to the reference orientation. In the outer loop, the desired thrust force is determined using estimates of the aerodynamic forces and an MRAC law. The reference orientation is determined as a solution of the orthogonal Procrustes' problem, which finds the smallest rotation from the current orientation of the thrust force, to the orientation of th desired thrust force. The angular velocity and acceleration cannot be deduced by taking the time derivative of the solution of the orthogonal Procrustes' problem due to the discontinuous nature of the singular value decomposition. Therefore, the twice continuously differentiable function, spherical linear interpolation, is used to find a geodesic joining the unit quaternion capturing the vehicle's current orientation, and the unit quaternion capturing the reference orientation. An interesting result is that the angular velocity and acceleration depend only on the first and second derivatives of the scalar-valued function which parameterizes the spherical linear interpolation function; the actual function is immaterial. However, determining the shape of this function is nontrivial, and hence, an approach inspired by model predictive control is used. In the inner loop, estimates of the aerodynamic moment and an MRAC law are used to stabilize the rotational dynamics, and the thrust force is allocated to the individual propellers. The validity of the proposed control scheme is presented in simulation. An integrated guidance and control system for autonomous UAS is proposed to maneuver in an unknown, dynamic, and potentially hostile environment in a reckless or tactical manner as prescribed by the user. Tactical maneuvering in this guidance and control system is enabled through exploitation of obstacles in the environment for shelter as the vehicle approaches its goal. Reckless maneuvering is enabled by ignoring the presence of nearby obstacles while proceeding towards the goal, while remaining collision-free. The demarcation of reckless and tactical behaviors are bio-inspired, since these tactics are used by animals or ground-based troops. The guidance system fuses a path planner, collision-avoidance algorithm, vision-based navigation system, and a trajectory planner. The path planner is based on the A* search algorithm, and a custom tunable cost-to-come and heuristic function are proposed to enable the exploitation of the obstacles' set for shelter by decreasing the weight of edges in the underlying graph that capture nodes close to the obstacles' set. The consistency of the heuristic is established, and hence, the search algorithm will return an optimal solution, and not expand nodes multiple times. In realistic scenarios, fast replanning is necessary to ensure that the system realizes the desired behavior, and does not collide with obstacles. The trajectory planner is based on fast model predictive control (fMPC), and thus, can be executed in real time. A custom tunable cost function, which weighs the importance of proximity to the obstacles' set and proximity to the goal, is employed to provide another mechanism for enabling tactical behaviors. The novel collision avoidance algorithm is based on the solution of a particular class of semidefinite programming problems, that is, quadratic discrimination. The collision avoidance algorithm produces convex sets of free space near the UAS by finding ellipsoids that separate the UAS from the obstacles' set. The convex sets are used in the fMPC framework as inequality constraints. The collision avoidance algorithm's computational burden is determined empirically, and is shown to be faster than two similar algorithms in the literature. The modules above are integrated into a single guidance system, which supplies reference trajectories to an arbitrary control system, and the validity of the proposed approach is exhibited in several simulations and flight tests. Furthermore, a taxonomy of flight behaviors is presented to understand how the tunable parameters affect the recklessness or stealthiness of the resulting trajectory. Lastly, an integrated guidance and control system for autonomous UAS performing tactical coverage in an unknown, dynamic, and potentially hostile environment in a reckless or tactical manner as prescribed by the user is presented. The guidance problem for coverage concerns strategies and route planning for gathering information about an environment. The aim of gathering information about an unknown environment is to aid in situational awareness and planning for service organizations and first-responders. To address this problem, goal selection, path planning, collision avoidance, and trajectory planning are integrated. A novel goal selection algorithm based on the Octree data structure is proposed to autonomously determine goal points for the path planner. In this algorithm, voxel maps deduced by a navigation system, which capture the occupancy and exploration status of areas of the environment, are segmented into partitions that capture large unexplored areas, and large explored areas. Large unexplored areas are used as candidates for goal points. The feasibility of goal points is determined by employing a greedy $A^*$ technique. The algorithm boasts tunable parameters that allow the user to specify a greedy or systematic behavior when determining a sequence of goal points. The computational burden of this technique is determined empirically, and is shown to be useful for real-time use in realistic scenarios. The path planner is based on the Lifelong Planning $A^*$ ($LPA^*$) search algorithm which is shown to have advantages over the $A^*$ technique. A custom tunable cost-to-come and heuristic function are proposed to enable tactical or reckless path planning. A novel collision avoidance algorithm is proposed as an improved version of the aforementioned collision avoidance algorithm, where the volume of the resulting constraint sets are improved, and thus, more of the free space is captured by the convex set, and hence, the trajectory planner can exploit more of the environment for tactical maneuvering. This algorithm is based on semidefinite programming and a fast approximate convex hull algorithm. The trajectory planner is based on fMPC, employs a custom cost function to enable tactical maneuvering by coasting the surface of obstacles and regulation of the desired acceleration as a function of proximity to shelter, employs barrier functions to constrain the attitude of the vehicle and ensure thrust positivity, and employs a quadrotor UAS' output feedback linearized equations of motion as differential constraints to enable aggressive maneuvering. The efficacy of the proposed system is validated using a custom-made C++ simulator. / Doctor of Philosophy / In recent years, unmanned aerial systems (UAS) such as quadcopters, hexacopters, and octocopters, have seen increased popularity for a myriad of applications including crop monitoring, photography, surveying, surveillance, wireless network extension, search and rescue, firefighter support, and military operations, to name a few. This list of applications stems from UAS' maneuverability, adaptability, accessibility, and their absence of an onboard pilot. While some of these applications can be executed with current capabilities, the performance of these systems could be improved, and there are many applications where UAS could be used to fulfil substantial roles in areas such as logistics, tactical surveillance, and direct human-interaction. However, these applications require a considerable improvement in autopilot design for UAS; shortcomings of current capabilities are identified in this thesis. Indeed, one of the most important improvements to be made is enabling fully autonomous operations where limited human intervention and oversight is necessary for mission success. In this thesis, we present two adaptive control systems for tailsitter UAS to enable accurate trajectory tracking in realistic scenarios with degraded conditions, such as inclement weather with unsteady winds, poorly-modeled dynamics as a result of negligence or a cost-benefit analysis, failing actuators due to overuse or damage from collisions. In the first adaptive control system, we focus on the tailsitter UAS' longitudinal dynamics, and employ a novel adaptive control technique to stabilize the system. In the second adaptive control system, we do not separate the longitudinal and lateral-directional dynamics, and split the tailsitter UAS' translational and rotational dynamics into outer and inner loops, respectively. In this control system, the windup problem is addressed using barrier functions, the reference orientation is determined as a solution to the orthogonal Procrustes' problem, and the system's dynamics are stabilized using model reference adaptive control. Furthermore, in this dissertation, we develop and present a guidance and control system which can be used to enable autonomous intelligence, surveillance, reconnaissance, and logistics (ISRL) operations in unknown, dynamic, and potentially hostile environments. The guidance system enables the UAS to achieve a user-defined behavior which ranges from tactical to reckless. The tactical or reckless behaviors are enabled through the guidance system's path planner, which is based on the A* search algorithm employing custom cost and heuristic function. Similarly, the guidance system's trajectory planner, which is based on fast model predictive control (fMPC), enables tactical or reckless behaviors through a custom cost function. The problem of collision-avoidance is handled through the path planner, which returns collision-free paths, and a novel constraint set generation algorithm which deduces regions of free space near the UAS; these regions are used as constraint sets for the trajectory planner. We validate the proposed approach in simulation and flight tests, and present a taxonomy of flight behaviors.

Un des buts de la recherche et du développement des systèmes de transport intelligents est d'augmenter le confort et la sécurité des passagers, tout en réduisant la consommation d'énergie, la pollution de l'air et le temps de déplacement. L'introduction de voitures complètement autonomes sur la voie publique nécessite la résolution d'un certain nombre de problèmes techniques et en particulier de disposer de modules de planification de trajectoire robustes. Ce travail de thèse s'inscrit dans ce cadre. Il propose une architecture modulaire pour la planification de trajectoire d'un véhicule autonome. La méthode permet de générer des trajectoires constituées de courbes interpolées adaptées à des environnements complexes comme des virages, des ronds-points, etc., tout en garantissant la sécurité et le confort des passagers. La prise en compte de l'incertitude des systèmes de perception, des limites physiques du véhicule, de la disposition des routes et des règles de circulation est aussi assurée dans le calcul de la trajectoire. L'algorithme est capable de modifier en temps réel la trajectoire prédéfinie de façon à éviter les collisions. Le calcul de la nouvelle trajectoire maintient les accélérations latérales à leur minimum, assurant ainsi le confort du passager. L'approche proposée a été évaluée et validée dans des environnements simulés et sur des véhicules réels. Cette méthode permet d'éviter les obstacles statiques et dynamiques repérés par le système de perception.Un système d'aide à la conduite pour le contrôle partagé basé sur cette architecture est introduit. Il prend en compte l'arbitrage, la surveillance et le partage de la conduite tout en maintenant le conducteur dans la boucle de contrôle. Il laisse le conducteur agir tant qu'il n'y a pas de danger et interagit avec le conducteur dans le cas contraire. L'algorithme se décompose donc en deux processus : 1) évaluation du risque et, s'il y a un risque avéré 2) partage du contrôle à l'aide de signaux haptiques via le volant.La méthode de planification de trajectoire présentée dans cette thèse est modulaire et générique. Elle peut être intégrée facilement dans toute architecture d'un véhicule autonome. / Developments in the Intelligent Transportation Systems (ITS) field show promising results at increasing passengers comfort and safety, while decreasing energy consumption, emissions and travel time. In road transportation, the appearance of automated vehicles is significantly aiding drivers by reducing some driving-associated tedious tasks. However, there is still a long way to go before making the transition between automated vehicles (i.e. vehicles with some automated features) and autonomous vehicles on public roads (i.e. fully autonomous driving), specially from the motion planning point of view. With this in mind, the present PhD thesis proposes the design of a generic modular architecture for automated vehicles motion planning. It implements and improves curve interpolation techniques in the motion planning literature by including comfort as the main design parameter, addressing complex environments such as turns, intersections and roundabouts. It will be able to generate suitable trajectories that consider measurements' incertitude from the perception system, vehicle’s physical limits, the road layout and traffic rules. In case future collision states are detected, the proposed approach is able to change---in real-time---the current trajectory and avoid the obstacle in front. It permits to avoid obstacles in conflict with the current trajectory of the ego-vehicle, considering comfort limits and developing a new trajectory that keeps lateral accelerations at its minimum. The proposed approach is tested in simulated and real urban environments, including turns and two-lane roundabouts with different radii. Static and dynamic obstacles are considered as to face and interact with other road actors, avoiding collisions when detected. The functional architecture is also tested in shared control and arbitration applications, focusing in keeping the driver in the control loop to addition the system's supervision over drivers’ knowledge and skills in the driving task. The control sharing advanced driver assistance system (ADAS) is proposed in two steps: 1) risk assessment of the situation in hand, based on the optimal trajectory and driving boundaries identified by the motion planning architecture and; 2) control sharing via haptic signals sent to the driver through the steering wheel. The approach demonstrates the modularity of the functional architecture as it proposes a general solution for some of today's unsolved challenges in the automated driving field.

Esta tese trata de um robô bípede em caminhar dinâmico. Neste robô, que normalmente é um sistema sub-atuado, fazemos uso de um disco de inércia que funciona num certo sentido como um atuador adicional. Através deste disco, obtém-se mais liberdade para a elaboração de passos repetitivos e um aumento na robustez. Por outro lado, o sistema de controle dos passos deve controlar, além do passo propriamente dito, também a velocidade do disco, de modo que não sejam saturados os atuadores (motores elétricos). Apresentamos então um controlador capaz de realizar estas ações simultaneamente. / This Thesis is about a bipedal robot in a dynamic walking gait. In this robot, which is usually a under-actuated system, a inertial wheel is employed and acts as an additional actuator. By using this wheel, one can design a cyclic walking gait with increased robustness and with more freedom. On the other hand, the control system must take care of the step itself, and also must ensure that the wheel speed does not exceed the actuators (motors) limits. We present a controller able to perform this tasks.

presente trabalho consolida um estudo para o planejamento de trajetória livre de colisão para um robô pneumático com 5 graus de liberdade aplicando três algoritmos metaheurísticos: algoritmos metaheurísticos por vagalumes, algoritmos metaheurísticos por enxames de partículas e algoritmos genéticos. No que se refere à aplicação de algoritmos metaheurísticos ao estudo de planejamento de trajetória de robôs manipuladores na presença de obstáculos, existem diferentes tipos de técnicas para evitar colisões que consideram os efeitos cinemáticos e dinâmicos na obtenção de trajetórias com o menor tempo, torque, etc. Neste estudo, são propostas contribuições à aplicação dessas técnicas especificamente a robôs manipuladores pneumáticos, sobretudo, no que diz respeito às características específicas dos servoposicionadores pneumáticos, como, por exemplo, a modelagem do atrito desses sistemas, o cálculo da massa equivalente, etc. A metodologia utilizada é definida em duas etapas. A primeira delas consiste na obtenção de pontos intermediários, adquiridos considerando a menor distância entre os mesmos e o ponto final, gerados considerando a presença de obstáculos (cilindros, cubos e esferas) Esses obstáculos são mapeados em regiões de colisão, que constituem restrições para o problema de otimização. A segunda etapa baseia-se no estudo do planejamento de trajetórias: aplicam-se b-splines de 5º e 7º grau na interpolação dos pontos intermediários, com vistas à obtenção de trajetórias que considerem, de um lado, a menor força dos atuadores associada à dinâmica do manipulador em estudo e, de outro, restrições cinemáticas e dinâmicas, determinadas por meio das características operacionais dos servoposicionadores pneumáticos. Os resultados mostram que a metodologia proposta é adequada para tarefas de manipulação de peças na presença de obstáculos, uma vez que os pontos intermediários situam-se fora da região de colisão nos três casos aqui apresentados. Além disso, quanto à segunda etapa, observou-se que as trajetórias de 5º e 7º grau apresentaram resultados similares, de maneira que os erros obtidos poderiam ser melhorados analisando aspectos associados ao controlador do robô em estudo. / The thesis presents a study for collision-free trajectory planning for a pneumatic robot with 5 degrees of freedom applying three metaheuristic algorithms: firefly metaheuristic algorithm, particle swarm optimization and genetic algorithms. As regards the application of metaheuristic algorithms to the study of the trajectory planning of manipulating robots in the presence of obstacles, there are different types of techniques to avoid collisions that consider the kinematic and dynamic effects, obtaining trajectories with the optimal time, torque, etc. In this study, contributions are made to the application of these techniques specifically to pneumatic manipulator robots, particularly with regard to the specific characteristics of pneumatic servo-actuators, such as friction modeling of these systems, calculation of equivalent mass, etc. The methodology used is defined in two steps. The first one consists of obtaining intermediate points, acquired considering the smallest distance between the intermediate points and the final point, generated considering the presence of obstacles (cylinders, cubes and spheres) These obstacles are mapped in collision regions, which are constraints to the optimization problem. The second step is based on the study of the trajectory planning: 5th and 7th degree b-splines are applied in the interpolation of the intermediate points, in order to obtain trajectories that consider the smallest actuator force associated to the dynamics of the manipulator and the kinematic and dynamic constraints, determined by the operational characteristics of pneumatic servo-positioners. The results show that the proposed methodology is suitable for tasks of manipulating parts in the presence of obstacles because the intermediate points are outside the collision region in the three cases presented here. In addition, it was observed that the trajectories of 5th and 7th degree presented similar results, so that the errors obtained could be improved by analyzing aspects associated to the controller of the robot.

Este trabalho trata do planejamento de trajetórias que minimizam as perdas elétricas no KA\'I yxo, um robô escalador de árvores que tem por finalidade realizar monitoramento ambiental em florestas através da coleta de diferentes tipos de dados. Como essa aplicação requer que o robô permaneça em ambientes remotos, o estudo de técnicas que reduzam as perdas de energia a fim de que se aumente o tempo em operação do robô se mostra relevante, sendo a minimização das perdas elétricas uma contribuição importante nesse sentido. Estruturalmente, o KA\'I yxo consiste em um robô bípede com duas garras e quatro ligamentos interconectados por três juntas rotacionais. Além disso, seu mecanismo de andadura foi biologicamente inspirado na forma de locomoção observada em lagartas mede-palmos, o que permitiu tratar o robô como um manipulador industrial, cuja base é o ligamento associado à garra engastada e cujo efetuador é o ligamento associado à garra livre. Com isso, quando conveniente, o robô foi tratado em dois casos, conforme a garra que se encontra engastada. Inicialmente, realizou-se a modelagem matemática do robô, obtendo-se as equações cinemáticas direta e inversa, e dinâmicas, bem como o modelo das juntas segundo a abordagem do controle independente por junta. Posteriormente, formulou-se um problema de controle ótimo, solucionado através de um método numérico que o transformou em um problema de programação quadrática, que por sua vez foi resolvido iterativamente. Por fim, as trajetórias ótimas planejadas foram implementadas no robô real e, como forma de validação, as novas perdas elétricas foram comparadas com as das trajetórias anteriormente executadas pelo robô, determinando-se a correspondente economia de energia. / This work deals with the minimum-energy trajectory planning, related to the electrical losses, in KA\'I yxo, a tree-climbing robot that aims to perform environmental monitoring in forests through the collection of different types of data. As this application requires that the robot remains in remote environments, the study of techniques that reduce energy losses in order to increase the operation time of the robot is shown to be relevant, and the minimization of the electrical losses is an important contribution in this sense. Structurally, KA\'I yxo consists of a biped robot with two claws and four links interconnected by three revolute joints. In addition, its gait mechanism was biologically inspired in the form of locomotion observed in caterpillars, allowing to treat the robot as an industrial manipulator, which base is the link associated with the fixed claw and which end-effector is the link associated with the free claw. In consequence, when convenient, the robot was treated in two cases, according to the claw that is fixed. Initially, the mathematical model of the robot was developed, being obtained the forward and inverse kinematic and dynamic equations, as well as the model of the joints according to the independent joint control approach. Subsequently, an optimal control problem was formulated, which was solved through a numerical method that turned it into a quadratic programming problem, which in turn was solved iteratively. Finally, the planned optimal trajectories were implemented in the real robot and, as a form of validation, the new electrical losses were compared with those of the trajectories previously executed by the robot, being determined the corresponding energy saving.

Nos últimos anos, a área de robôs aéreos vêm se tornando cada vez mais importantes no dia-a-dia, em diversos usos, em que se pode destacar: segurança pública e particular, agricultura de precisão, registro fotográfico de eventos, serviços de entregas e apoio a diversas outras áreas, como monitoramento ambiental.Para que tais robôs aéreos possam cumprir suas mais variadas tarefas, faz-se necessária uma etapa de planejamento de movimento, que consiste em encontrar um caminho factível entre a postura atual e uma postura final (também chamada de postura alvo ou meta) do robô aéreo. Neste trabalho, a tarefa de planejamento de movimento é abordada para o caso tridimensional em ambientes dinâmicos, nos quais não se assume que todos os obstáculos permanecerão fixos ao longo do trajeto. Derivado da Teoria de Campos Potenciais Harmônicos, os Campos Potenciais Modificados (CPM) permitem a distorção do campo potencial favorecendo uma direção específica de chegada a postura meta. Tais CPM resultam em um planejador de movimentos para ambientes dinâmicos e multidimensionais, em especial, o caso tridimensional. / In recent years, aerial robots have become increasingly important in day-to-day situations, in several uses, in which we can highlight: public and private security, precision agriculture, photographic record of events, delivery and support to several other areas, such as environmental monitoring. In order for aerial robots perform their broad range of tasks, a motion planning step is necessary. Motion planning consists in finding a feasible path between the current posture and a final posture (also called target posture or goal) of a robot. In this work, the task of motion planning is approached in three-dimensional and dynamic environments, in which it is not assumed that all the obstacles will remain fixed along the trajectory. Derived from Harmonic Potential Field Theory, Modified Potential Fields (MPF) allows a controlled distortion of the potential field, as an example, towards a specific direction of arrival to the target posture. Such MPF results in a motion planner for dynamic and multidimensional environments, especially the three-dimensional case.