This thesis describes work in a variety of topics related to aerial robotics, including system identification, state estimation, control, and path planning. The path planners described in this thesis are used to guide a fixed-wing UAV along paths that optimize the aircraft's ability to track a ground target. Existing path planners in the literature either ignore occlusions entirely, or they have limited capability to handle different types of paths. The planners described in this thesis are novel in that they specifically account for the effect of occlusions in urban environments, and they can produce a much richer set of paths than existing planners that account for occlusions. A 3D camera positioning system from Motion Analysis is also described in the context of state estimation, system identification, and control of small unmanned rotorcraft. Specifically, the camera positioning system is integrated inside a control architecture that allows a quadrotor helicopter to fly autonomously using truth data from the positioning system. This thesis describes the system architecture in addition to experiments in state estimation, control, and system identification. There are subtleties involved in using accelerometers for state estimation onboard flying rotorcraft that are often ignored even by researchers well-acquainted with the UAV field. In this thesis, accelerometer-rotorcraft behavior is described in detail. The consequences of ignoring accelerometer-rotorcraft behavior are evaluated, and an observer is presented that achieves better performance by specifically modeling actual accelerometer behavior. The observer is implemented in hardware and results are presented.

Tato disertační práce se zabývá metodami modelování elektronického zařízení letadel. První část je stručným přehledem klasických metod modelování systémů a adaptivních, fuzzy a hybridních metod používaných převážně k black-box modelování. Cílem práce je vytvořit algoritmus pro identifikaci a modelování obecného systému, který může být nelineární, dynamický a velmi složitý, například co do množství rozměrů. Předpokládá se, že model má několik vstupů a výstupů. V hlavní části práce je rozebrána metoda, která patří mezi hybridní systémy, protože kombinuje fuzzy systém s parametricky definovanými pravidly a regresní neuronovou síť. Nejprve je zmíněn základní princip regresní sítě a způsob určení jejího parametru strmosti, dále se kapitola zabývá zavedením fuzzy pravidel do této sítě. Třetí část se zabývá jedním z hlavních bodů práce, paralelními výpočty. Výsledný algoritmus je navržen pro paralelní zpracování, protože výpočetní čas může být v případě složitějších modelů příliš vysoký, případně nelze výsledky získané ze sítě vyhodnotit pomocí výpočtu v jednom vlákně. V závěru práce je metoda ověřena na datech získaných z měření zmenšeného modelu letadla. Ověření je provedeno pomocí střední kvadratické odchylky a srovnáním s odpovídajícím modelem vytvořeným pomocí vícevrstvé neuronové sítě trénované zpětným šířením chyby s algoritmem Levenberg-Marquardt.

The paper presents design solution and physical implementation of a system for examination of electro hydraulic steering based on OSPE 200 components. The implementation is based on synthesis of required hydraulic and structure parameters, presented in a previous paper. Now we present the interconnection of the digital control system and the closed-loop flow diagram. A formal description of embedded software is presented too, which supports operation of PI control algorithm in real-time. Identification is performed based on experimentally reported the transitional process by developing mathematical models. Presents the structure and capabilities of the models for identification, as well as procedures for their validation.

System identification concerns how to construct mathematical models of dynamic systems based on experimental data. A very important application of system identification is in model-based control design. In such applications it is often possible to externally excite the system during the data collection experiment. The properties of the exciting input signal influence the quality of the identified model, and well-designed input signals can reduce both the experimental time and effort. The objective of this thesis is to develop algorithms and theory for minimum cost experiment design for system identification while guaranteeing that the estimated model results in an acceptable control performance. We will use the framework of application oriented Optimal Input Design (OID). First, we study how to find a convex approximation of the set of models that results in acceptable control performance. The main contribution is analytical methods to determine application sets for controllers with no explicit control law, for instance Model Predictive Control (MPC). The application oriented OID problem is then formulated in time domain to enable the handling of signals constraints, which often comes from the physical limitations on the plant and actuators. The framework is the extended to closed-loopsystems. Here two different cases are considered. The first case assumes that the plant is controlled by a general (either linear or non-linear) but known controller. The main contribution here is a method to design an external stationary signal via graph theory such that the identification requirements and signal constraints are satisfied. In the second case application oriented OID problem is studied for MPC. The proposed approach here is a modification of a results where the experiment design requirements are integrated to the MPC as a constraint. The main idea is to back off from the identification requirements when the control requirements are violating from their acceptable bounds. We evaluate the effectiveness of all the proposed algorithms by several simulation examples. / <p>QC 20150126</p>

The subject of UCAV design is an important topic nowadays and many countries have their own programmes. An international group, under the initiative of the NATO RTO AVT-201 Task group, titled “Extended Assessment of Reliable Stability &amp; Control Prediction Methods for NATO Air Vehicles”, is currently performing intensive analysis on a generic UCAV configuration, named SACCON. In this thesis the stability and control characteristics of the SACCON are investigated, with the purpose of carrying out a comprehensive assessment of the flying qualities of the design. The study included the generation of the complete aerodynamic database of the aircraft, on the basis of the experimental data measured during TN2514 and TN2540 campaigns at DNW-NWB low speed wind tunnel. Moreover, system identification techniques were adopted for the extraction of dynamic derivatives from the time histories of forced oscillation runs. The trim of the aircraft was evaluated across the points of a reasonable test envelope, so as to define a set of plausible operative conditions, representing the reference conditions for subsequent linearization of the dynamic model. The study provided a thorough description of the stability and control characteristics and flying qualities of the unaugmented SACCON, laying the groundwork for future improvement and validation of the configuration in the next design stages.

The market for electric vehicles (EV) is growing rapidly. The rise of EVs is most prominent for passenger vehicles, but trucks and busses are also quickly becoming electrified. Scania aims to be the leader of this transition. A central part of the EV is the Lithium-ion battery. In order to use the battery in the most efficient manner a Battery Management System (BMS) is needed. A key part of the BMS is a model that describes the battery as a system where the input is the current and the output is the terminal voltage. The dynamics of the battery is affected by external factors, called scheduling variables, that should be taken in to account in order to acquire an accurate model. This thesis aims to capture this behavior by the identification of a Linear Parameter Varying (LPV) model that has State of Charge (SOC) and temperature as scheduling variables. The LPV model was identified by first performing a set of local system identifications at varying levels of the scheduling variables. From this, a set of different LPV model structures were set up and then optimized with the use of datasets with a wider coverage of the scheduling variables. The results showed that there are clear advantages in using an LPV model compared to a traditional constant model, but that the robustness of the model largely is dependent on the choice of the data used for optimization.

The demand of space applications has been increasing over the years. This has derivedin new satellites structures that required from precise and robust control management.The satellite design is evolving towards the development of lighter structures. The combinationof lighter structures with precise and robust control has arisen the problem ofstructure vibration control. The control design of structures with large appendages likeantennas, booms or solar panels has become a challenge. The flexible dynamics of theappendages needs to be considered when performing the attitude analysis of the satellite,since these parts can be easily excited by the environment perturbations such us gravity,gravity gradient or solar wind. The objective of these research project is to develop ahigh-fidelity model plant of a satellite with flexible panels and review different systemidentification techniques used to observe the states of the system. The equations of themodel are reviewed and the model is verified against a multi-body software, Adams. Thesensors and actuators are selected and modelled for the control of the rigid body and theobservation of the rigid and flexible body. For the implementation of the flexible structureobservations a technique based in Genetic Algorithm is applied for optimal sensor location.Finally, different system identification techniques are reviewed for the identificationof modal parameters and rigid body parameters. The results illustrate the performanceof the model and how the different system identification techniques are performed whenobserving the model states.

京都大学 / 新制・課程博士 / 博士(情報学) / 甲第23321号 / 情博第757号 / 新制||情||129(附属図書館) / 京都大学大学院情報学研究科数理工学専攻 / (主査)教授 太田 快人, 教授 山下 信雄, 教授 大塚 敏之 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DGAM

Empirical models based on linear and nonlinear potential theory that determine the forces on Wave Energy Converters (WECs) are essential as they can be used for structural, mechanical and control system design as well as performance prediction. In contrast to empirical modelling, Computational Fluid Dynamics (CFD) solves the mass and momentum balance equations for fluid domains. CFD and linear potential theory models represent two extreme in terms of capturing the full range of hydrodynamic effects. These are classified as white box models and the structure of these models is completely derived from first principles understanding of the system. In contrast black box models like a Artificial Neural Networks and Auto-Regressive with, Exogenous Input (ARX), map input and output behaviour of a system without any specific physics based structure. Grey box models do not strictly follow a first principles approach but are based on some observations of relationships between the hydrodynamic effects (e.g. buoyancy force) and system state (e.g. free surface height). The objective of this thesis is to propose a data driven grey box modelling approach, which is computationally efficient compared to high fidelity white box mod- els and still sufficiently accurate for the purpose of determining hydrodynamic forces on heaving WECs. In this thesis, a unique data driven approach that combines features from existing works in modelling of WEC and application of nonlinear hysteretic systems is developed. To that end a CFD based Numerical Wave Tank that could provide the data needed to populate the new modelling framework is used. A hull which hydrodynamically represents a Self Reacting Point Absorbers (SRPAs) with heave plate is subjected to pan-chromatic wave fields and is forced to oscillate concomitantly. The results provide evidence that a SRPA with heave plate exhibits nonlinear relationships with motion parameters including relative position, velocity and acceleration. These parameters show causal relationships with the hydrodynamic force. A simulation methodology to establish confidence in the components of a model framework is developed and the hydrodynamic forces on SRPAs with heave plate and bulbous tank have been analyzed and compared. Two sets of numerical simulation were conducted. Firstly, the WECs were restricted to all degrees of freedom and subjected to monochromatic waves and later the WECs were oscillated at various frequency in a quiescent numerical tank. These results were validated against existing experimental data. Earlier attempts by other authors to develop a data-driven model were limited to simple hulls and did not include rate dependent nonlinearities that develop for heave plates. These studies laid the foundation to current work. The model framework developed in this thesis accounts for the nonlinear relationship between force and parameters like velocity and acceleration along with hysteretic relationship between force and velocity. This modelling framework has a nonlinear static, a hysteresis (Bouc-Wen model) and a dynamic (ARX model) block. In this work the Bouc-Wen model is employed to model the hysteresis effect. Five different models developed from this modelling framework are analyzed; two are state dependent models, while the other three required training to identify dynamic order of model equations. These latter models (Hammerstein, rate dependent Hammerstein and rate dependent KGP models) have been trained and validated for various cases of fixed and oscillating HP cylinder. The results demonstrate significant improvement (max 39%) in prediction accuracy of hydrodynamic forces on a WEC with heave plate, for the model in which a rate dependent hysteresis block is coupled with Hammerstein or KGP models. / Graduate

This dissertation addresses motion planning, modeling, and feedback control for autonomous vehicle systems. A hierarchical approach for motion planning and control of nonlinear systems operating in obstacle environments is presented. To reduce computation time during the motion planning process, dynamically feasible trajectories are generated in real-time through concatenation of pre-specified motion primitives. The motion planning task is posed as a search over a directed graph, and the applicability of informed graph search techniques is investigated. Specifically, a locally greedy algorithm with effective backtracking ability is developed and compared to weighted A* search. The greedy algorithm shows an advantage with respect to solution cost and computation time when larger motion primitive libraries that do not operate on a regular state lattice are utilized. Linearization of the nonlinear system equations about the motion primitive library results in a hybrid linear time-varying model, and an optimal control algorithm using the L2-induced norm as the performance measure is applied to ensure that the system tracks the desired trajectory. The ability of the resulting controller to closely track the trajectory obtained from the motion planner, despite various disturbances and uncertainties, is demonstrated through simulation. Additionally, an approach for obtaining dynamically feasible reference trajectories and feedback controllers for a small unmanned aerial vehicle (UAV) based on an aerodynamic model derived from flight tests is presented. The modeling approach utilizes the two step method (TSM) with stepwise multiple regression to determine relevant explanatory terms for the aerodynamic models. Dynamically feasible trajectories are then obtained through the solution of an optimal control problem using pseudospectral optimal control software. Discrete-time feedback controllers are then obtained to regulate the vehicle along the desired reference trajectory. Simulations in a realistic operational environment as well as flight testing with the feedback controller demonstrate the capabilities of the approach. The TSM is also applied for system identification of an aircraft using motion capture data. In this application, time domain system identification techniques are used to identify both linear and nonlinear aerodynamic models of large-amplitude pitching motions driven by control surface deflections. The resulting models are assessed based on both their predictive capabilities as well as simulation results. / Ph. D.