Heavy load equipment, such as tractors, shovels, cranes, airplanes, etc, often employ fluid power (i.e. hydraulic) systems to control their loads by way of valve adjustment in a pump-valve control configuration. Most of these systems have low energy efficiency as a consequence of pressure losses across throttle valves. Much of the energy is converted into heat energy which can have determinantal effects on component life and the surrounding environment. From an energy efficiency point of view, an ideal hydraulic system is one that does not include any throttling valve. One such circuit is made of a variable pump and motor load (pump/motor configuration). The velocity of the load is controlled by manipulating the pump displacement or by changing the rotary speed of the pump shaft. In such a system, the transient response of the load is often unsatisfactory because it is difficult to quickly and accurately manipulate the pump displacement or change shaft speed. Thus circuit design must be a compromise between the energy efficiency of the pump/motor system and the controllability of a pump/valve/motor combination. One possible compromise is to use a pump-valve configuration which reduces energy losses across the valve. One way to achieve this is by controlling the pressure drop across the valve and limiting it to a small value, independent of load pressure. Based on this idea, a type of hydraulic control system, usually called load-sensing (LS), has recently been used in the flow power area. This type of system, however, is complex and under certain operating conditions exhibits instability problems. Methods for compensating these instabilities are usually based on a trial-and-error approach. Although some research has resulted in the definition of some instability criterion, a comprehensive and verifiable approach is still lacking. This research concentrates on identifying the relationship between system parameters and instability in one particular type of LS system. Due to the high degree of non-linearity in LS systems, the instabilities are dependent on the steady state operating point. The study therefore concentrates first on identifying all of the steady state operating points and then classifying them into three steady state operating regions. A dynamic model for each operating region is developed to predict the presence of instabilities. Each model is then validated experimentally. This procedure, used in the study of the LS system, is also applied to a pressure compensated (PC) valve. A PC valve is one in which the flow rate is independent in variations to load pressure. A system which combines a LS pump and a PC valve (for the controlling orifice) is called a load sensing pressure compensated (LSPC) system. This research, then, examines the dynamic performance of the LSPC system using the operating points and steady state operating regions identified in the first part of the research. The original contributions of this research include: (a) establishment of three steady state operating conditions defined as Condition I, II & III, which are based on the solution of steady state non-linear equations; (b) the provision of an empirical model of the orifice discharge coefficient suitable for laminar and turbulent flow, and the transition region between them; (c) and the development of an analytical expression for orifice flow which makes it possible to accurately model and simulate a hydraulic system with pilot stage valve or pump/motor compensator. These contributions result in a practical and reliable method to determine the stability of a LS or LSPC system at any operating point and to optimize the design of the LS or LSPC system.

Aquesta Tesi Doctoral se centra en el disseny d'un nou linealitzador de Predistorsió Digital (Digital Predistortion - DPD) capaç de compensar la dinàmica i els efectes no lineals introduïts pels Amplificadors de Potència (Power Amplifiers - PAs). Un dels trets més rellevants d'aquest nou predistorsionador digital i adaptatiu consisteix en ser deduïble a partir d'un model de PA anomenat Nonlinear Auto-Regressive Moving Average (NARMA). A més, la seva arquitectura multi-LUT (multi-Taula) permet la implementació en un dispositiu Field Programmable Gate Array (FPGA).La funció de predistorsió es realitza en banda base, per tant, és independent de la banda freqüencial on es durà a terme l'amplificació del senyal de RF, el que pot resultar útil si tenim en compte escenaris multibanda o reconfigurables. D'altra banda, el fet que aquest DPD tingui en compte els efectes de memòria introduïts pel PA, representa una clara millora de les prestacions aconseguides per un simple DPD sense memòria. En comparació amb d'altres DPDs basats en models més computacionalment complexos, com és el cas de les xarxes neuronals amb memòria (Time-Delayed Neural Networks - TDNN), la estructura recursiva del DPD proposat permet reduir el nombre de LUTs necessàries per compensar els efectes de memòria del PA. A més, la seva estructura multi-LUT permet l'escalabilitat, és a dir, activar or desactivar les LUTs que formen el DPD en funció de la dinàmica que presenti el PA.En una primera aproximació al disseny del DPD, és necessari identificar el model NARMA del PA. Un dels majors avantatges que presenta el model NARMA és la seva capacitat per trobar un compromís entre la fidelitat en l'estimació del PA i la complexitat computacional introduïda. Per reforçar aquest compromís, l' ús d'algoritmes heurístics de cerca, com són el Simulated Annealing o els Genetic Algorithms, s'utilitzen per trobar els retards que millor caracteritzen la memòria del PA i per tant, permeten la reducció del nombre de coeficients necessaris per caracteritzar-la. Tot i així, la naturalesa recursiva del model NARMA comporta que, de cara a garantir l'estabilitat final del DPD, cal dur a terme un estudi previ sobre l'estabilitat del model.Una vegada s'ha obtingut el model NARMA del PA i s'ha verificat l'estabilitat d'aquest, es procedeix a l'obtenció de la funció de predistorsió a través del mètode d'identificació predictiu. Aquest mètode es basa en la continua identificació del model NARMA del PA i posteriorment, a partir del model obtingut, es força al PA perquè es comporti de manera lineal. Per poder implementar la funció de predistorsió en la FPGA, cal primer expressar-la en forma de combinacions en paral·lel i cascada de les anomenades Cel·les Bàsiques de Predistorsió (BPCs), que són les unitats fonamentals que composen el DPD. Una BPC està formada per un multiplicador complex, un port RAM dual que actua com a LUT (taula de registres) i un calculador d'adreces. Les LUTs s'omplen tenint en compte una distribució uniforme dels continguts i l'indexat d'aquestes es duu a terme mitjançant el mòdul de l'envoltant del senyal. Finalment, l'adaptació del DPD consisteix en monitoritzar els senyals d'entrada i sortida del PA i anar duent a terme actualitzacions periòdiques del contingut de les LUTs que formen les BPCs. El procés d'adaptació del contingut de les LUTs es pot dur a terme en la mateixa FPGA encarregada de fer la funció de predistorsió, o de manera alternativa, pot ser duta a terme per un dispositiu extern (com per exemple un DSP - Digital Signal Processor) en una escala de temps més relaxada. Per validar l'exposició teòrica i provar el bon funcionalment del DPD proposat en aquesta Tesi, es proporcionen resultats tant de simulació com experimentals que reflecteixen els objectius assolits en la linealització del PA. A més, certes qüestions derivades de la implementació pràctica, tals com el consum de potència o la eficiència del PA, són també tractades amb detall. / This Ph.D. thesis addresses the design of a new Digital Predistortion (DPD) linearizer capable to compensate the unwanted nonlinear and dynamic behavior of power amplifiers (PAs). The distinctive characteristic of this new adaptive DPD is its deduction from a Nonlinear Auto Regressive Moving Average (NARMA) PA behavioral model and its particular multi look-up table (LUT) architecture that allows its implementation in a Field Programmable Gate Array (FPGA) device.The DPD linearizer presented in this thesis operates at baseband, thus becoming independent on the final RF frequency band and making it suitable for multiband or reconfigurable scenarios. Moreover, the proposed DPD takes into account PA memory effects compensation which representsan step forward in overcoming classical limitations of memoryless predistorters. Compared to more computational complex DPDs with dynamic compensation, such Time-Delayed Neural Networks (TDNN), this new DPD takes advantage of the recursive nature of the NARMA structure to relax the number of LUTs required to compensate memory effects in PAs. Furthermore, its parallel multi-LUT architecture is scalable, that is, permits enabling or disabling the contribution of specific LUTs depending on the dynamics presented by a particular PA.In a first approach, it is necessary to identify a NARMA PA behavioral model. The extraction of PA behavioral models for DPD linearization purposes is carried out by means of input and output complex envelope signal observations. One of the major advantages of the NARMA structure regards its capacity to deal with the existing trade-off between computational complexity and accuracy in PA behavioral modeling. To reinforce this compromise, heuristic search algorithms such the Simulated Annealing or Genetic Algorithms are utilized to find the best sparse delays that permit accurately reproducing the PA nonlinear dynamic behavior. However, due to the recursive nature of the NARMA model, an stability test becomes a previous requisite before advancing towards DPD linearization.Once the PA model is identified and its stability verified, the DPD function is extracted applying a predictive predistortion method. This identification method relies just on the PA NARMA model and consists in adaptively forcing the PA to behave as a linear device. Focusing in the DPD implementation, it is possible to map the predistortion function in a FPGA, but to fulfill this objective it is first necessary to express the predistortion function as a combined set of LUTs.In order to store the DPD function into a FPGA, it has to be stated in terms of parallel and cascade Basic Predistortion Cells (BPCs), which are the fundamental building blocks of the NARMA based DPD. A BPC is formed by a complex multiplier, a dual port RAM memory block acting as LUT and an address calculator. The LUT contents are filled following an uniform spacing procedure and its indexing is performed with the amplitude (modulus) of the signal's envelope.Finally, the DPD adaptation consists in monitoring the input-output data and performing frequent updates of the LUT contents that conform the BPCs. This adaptation process can be carried out in the same FPGA in charge of performing the DPD function, or alternatively can be performed by an external device (i.e. a DSP device) in a different time-scale than real-time operation.To support all the theoretical design and to prove the linearization performance achieved by this new DPD, simulation and experimental results are provided. Moreover, some issues derived from practical experimentation, such as power consumption and efficiency, are also reported and discussed within this thesis.

Nonlinear response history analysis is accepted as the most accurate analytical tool for seismic response determination. However, accurate estimation of displacement responses using conceptually simple, approximate analysis procedures is preferable, since there are shortcomings in the application of nonlinear response history analysis resulting from its complexity. An equivalent linearization procedure, which utilizes the familiar response spectrum analysis as the analysis tool and benefits from the capacity principles, is developed in this thesis study as an approximate method for predicting the inelastic seismic displacement response of MDOF systems under earthquake excitations. The procedure mainly consists of the construction of an equivalent linear system by reducing the stiffness of structural members which are expected to respond in the inelastic range. Different from similar studies in literature, equivalent damping is not explicitly employed in this study. Instead, predetermined spectral displacement demands are utilized in each mode of the equivalent linear system for the determination of global displacement demands. Response predictions of the equivalent linearization procedure are comparatively evaluated by using the benchmark nonlinear response history analysis results and other approximate methods including conventional pushover analysis and modal pushover analysis (MPA). It is observed that the proposed procedure results in similar accuracy with approximate methods which employ nonlinear analysis. Considering the conceptual simplicity of the procedure and the conventional analysis tools used in its application, presented equivalent linearization procedure can be suggested as a practically applicable method for the prediction of inelastic seismic displacement response parameters with sufficient accuracy.

In recent years, there has been a significant increase in the interest for snake like modular robots due to their superior locomotion capabilities in terms of versatility, adaptability and scalability. Passive wheeled planar snake like robots are a major category and they are being actively researched. Due to the nonholonomic constraints imposed on them, certain configurations lead to the singularity which must be avoided at all costs. Furthermore, it is vital to generate a locomotion pattern such that they can track a wide range of trajectories. All of these objectives must be accomplished smoothly and in an energy efficient manner. Studies indicate that meeting all of these requirements is a challenging problem. In this study, a novel form of the serpenoid curve is proposed in order to make the robot track arbitrary paths. A controller has been designed using the feedback linearization method. Afterwards, a new performance measure, considering both the efficiency and sustainability of the locomotion, has been proposed to evaluate the locomotion. Optimal parameters for the proposed serpenoid curve and the linear controller have been determined for efficient locomotion by running series of simulations. Relations between the locomotion performance, locomotion speed and eigenvalues of the linear controller have been demonstrated. Simulation results show striking differences between the locomotion by using the proposed serpenoid curve with optimal parameters and the locomotion by purely tracking a given path. Obtained results also indicate that the aforementioned requirements are met successfully and confirm the validity and consistency of the proposed performance measure.

This dissertation studies the design of an efficient adaptive digital baseband predistorter for modern cellular handsets that combines low power consumption, low implementation complexity, and high performance. The proposed enhancements are optimized for hardware implementation. We first present a thorough study of the optimal spacing of linearly-interpolated lookup table predistorters supported by theoretical calculations and extensive simulations. A constant-SNR compander that increases the predistorter's supported input dynamic range is derived. A corresponding low-complexity approximation that lends itself to efficient hardware design is also implemented in VHDL and synthesized with the Synopsys Design Compiler. This dissertation also proposes an LMS-based predistorter adaptation that is optimized for hardware implementation and compares the effectiveness of the direct and indirect learning architectures. A novel predistorter design with quadrature imbalance correction capability is developed and a corresponding adaptation scheme is proposed. This robust predistorter configuration is designed by combining linearization and I/Q imbalance correction into a single function with the same computational complexity as the widespread complex-gain predistorter.

Nature and its variety of motion forms have inspired new robot designs with inherentunderactuated dynamics. The fundamental characteristic of these controlled mechanicalsystems, called underactuated, is to have the number of actuators less than the number ofdegrees of freedom. The absence of full actuation brings challenges to planning feasibletrajectories and designing controllers. This is in contrast to classical fully-actuated robots.A particular problem that arises upon study of such systems is that of generating periodicmotions, which can be seen in various natural actions such as walking, running,hopping, dribbling a ball, etc. It is assumed that dynamics can be modeled by a classicalset of second-order nonlinear differential equations with impulse effects describing possibleinstantaneous impacts, such as the collision of the foot with the ground at heel strikein a walking gait. Hence, we arrive at creating periodic solutions in underactuated Euler-Lagrange systems with or without impulse effects. However, in the qualitative theory ofnonlinear dynamical systems, the problem of verifying existence of periodic trajectoriesis a rather nontrivial subject.The aim of this work is to propose systematic procedures to plan such motions and ananalytical technique to design orbitally stabilizing feedback controllers. We analyze andexemplify both cases, when the robotmodel is described just by continuous dynamics, andwhen continuous dynamics is interrupted from time to time by state-dependent updates.For trajectory planning, systems with one or two passive links are considered, forwhich conditions are derived to achieve periodicmotions by encoding synchronizedmovementsof all the degrees of freedom. For controller design we use an explicit form tolinearize dynamics transverse to the motion. This computation is valid for an arbitrarydegree of under-actuation. The linear system obtained, called transverse linearization, isused to analyze local properties in a vicinity of the motion, and also to design feedbackcontrollers. The theoretical background of these methods is presented, and developedin detail for some particular examples. They include the generation of oscillations forinverted pendulums, the analysis of human movements by captured motion data, and asystematic gait synthesis approach for a three-link biped walker with one actuator.

The aim of this thesis is to introduce a general framework for what is informally referred to as finite-time dynamics. Within this framework, we study hyperbolicity of reference trajectories, existence of invariant manifolds as well as normal hyperbolicity of invariant manifolds called Lagrangian Coherent Structures. We focus on a simple derivation of analytical results. At the same time, our approach together with the analytical results has strong impact on the numerical implementation by providing calculable expressions for known functions and continuity results that ensure robust computation. The main results of the thesis are robustness of finite-time hyperbolicity in a very general setting, finite-time analogues to classical linearization theorems, an approach to the computation of so-called growth rates and the generalization of the variational approach to Lagrangian Coherent Structures.

L'observateur est un outil essentiel pour accéder à l'information, non mesurable directement, d'un système dynamique. Dans le cas des systèmes non linéaires, il y a une grande carence concernant la synthèse d'observateur. Motivée par l'absence d'une solution générique, cette thèse élargit la classe des systèmes non linéaires pour lesquels on peut appliquer les observateurs connus. Dans l'approche adoptée, le système non linéaire est transformé à travers un changement de coordonnées sous forme normale d'observabilité qui a une structure adéquate à la synthèse d'observateurs. Les difficultés liées aux conditions d'existence d'un changement de coordonnées sont mises en évidence et des solutions sont proposées. Par conséquent, la classe des systèmes non linéaires qui peuvent se mettre sous une forme normale d'observabilité est élargie. Dans un premier temps, nous avons proposé une forme normale d'observabilité étendue dépendante de la sortie en augmentant l'espace d'état par des variables auxiliaires. Ainsi, nous avons établi les conditions nécessaires et suffisantes à l'existence d’un changement de coordonnées permettant d’obtenir une telle forme. En outre, nous avons proposé, pour certains modèles, des procédures heuristiques pour la mise sous forme normale d'observabilité étendue dépendante de la sortie. Dans un deuxième temps, nous avons traité la mise sous forme normale d'observabilité d'un système non linéaire partiellement observable. Enfin, nous avons abordé la transformation d'une classe spéciale de systèmes non linéaires sous la forme normale d'observabilité réduite. L'efficacité et l'intérêt des méthodes développées sont établis au travers de plusieurs applications. / Observer is an essential means to access to no-measurable information of a dynamical system. In the case of nonlinear systems, there is a great deficiency concerning the observer design theory. Motivated by the lack of a generic solution to observer design, this thesis enlarges the class of nonlinear systems which admit a standard observer. Using a geometrical approach, the considered nonlinear systems are transformed through a change of coordinates into observer form, which has an adequate structure to the observer design. The difficulties related to the conditions on the existence of such a change of coordinates are highlighted and solutions are proposed. Therefore, the class of nonlinear systems which can be transformed under an observer form is expanded. Firstly, we proposed an extended output depending observer form which does not preserve the size of the original state space and we established the sufficient conditions for the existence of a change of coordinates enabling to construct the proposed form. Moreover, we proposed a heuristic procedure to construct the extended output depending observer form of some models. Secondly, we dealt with the observer form for some partially observable nonlinear systems. Finally, we discussed the transformation into the reduced observer form for a class of nonlinear systems. The efficiency and interest of the developed methods is established through several applications.

En ingénierie sismique, il est admis que le comportement d’une structure soumise à de forts séismes soit caractérisé par des boucles d’hystérésis qui peuvent être amples ou étroites selon le type de structure impactée. La prise en compte de ce type de comportement non-linéaire dans un calcul temporel présente des difficultés liées à l’identification des paramètres, au coût numérique élevé, au risque de non-convergence. Dans ce contexte, la méthode de linéarisation équivalente, a été introduite en géotechnique dès les années 70. Elle reste peu utilisée dans le domaine des structures malgré les efforts de nombreux auteurs. Ce travail de thèse a pour objet l’étude du comportement linéaire équivalent dans le contexte des méthodes simplifiées d'évaluation de la réponse non-linéaire d'une structure en ingénierie sismique. Nous passons en revue les critères de linéarisation adoptés par les différentes méthodes qui recherchent l’équivalence (1) du déplacement maximum ou (2) de la quantité d’énergie dissipée ou (3) de la force de rappel. Nos analyses montrent que ces trois critères ne sont pas pertinents et/ou efficaces, conduisant à des méthodes peu robustes qui conduisent dans certains cas à des résultats inexplicables. Nous montrons le rôle important, négligé par toutes les méthodes disponibles, du contenu fréquentiel respectif des signaux et du système dans la détermination de la ductilité appelée. Sur cette constatation, nous introduisons une nouvelle méthode de linéarisation équivalente basée sur la fonction de transfert. Nous utilisons cette méthode pour explorer un plan d’expérience numérique dans lequel nous calculons les caractéristiques de fréquence et d’amortissement équivalents en fonction de la ductilité appelée pour différente configurations caractérisées par (a) le rapport entre fréquence de l’oscillateur et fréquence centrale du signal excitateur, (b) la pente d’écrouissage et (c) le modèle de comportement qui varie continument de élastoplastique à endommageant. Nous proposons deux nouvelles approches du comportement linéaire équivalent. La première, visant à améliorer la procédure statique non-linéaire de l’ATC40, utilise la rigidité sécante et le déplacement maximal. Elle fait intervenir une estimation de l’amortissement différente de celle de l’ATC40. Sa pertinence est établie par le fait qu’elle permet d’évaluer avec exactitude le déplacement maximal de systèmes canoniques non-linéaires. La seconde consiste à restituer la dynamique de la réponse d'un oscillateur non-linéaire au travers de la fonction de transfert. Sa pertinence est démontrée au travers des critères d’Anderson, avec notamment un critère relatif au spectre transféré. La détermination du comportement linéaire équivalent par fonction de transfert est validée sur des structures réelles au travers des essais sur voiles en béton armé (SAFE) et sur systèmes des tuyauteries (BARC et EPRI) / In earthquake engineering, it is common that the behaviour of a structure undergoing a strong motion is characterized by wise or narrow hysteresis loops depending on the type of behaviour of the structure. Considering this non-linear behaviour in a transient calculation requires a huge need of resources in terms of calculation time and memory. In this context, the method of equivalent linearization, consisting in the evaluation of the non-linear response of the structure has been introduced by geotechnical engineers In the 1970s. Despite efforts of many authors, this method is still not used in structural field. The goal of this research is to examine the linear equivalent behaviour in the context of the simplified method of evaluating the non-linear response of a structure in earthquake engineering. We review the criteria of equivalence adopted by many methods searching for the equivalence of (1) the maximum of displacement or (2) quantity of dissipated energy or (3) the restore force. Our argumentative analyses carry out that these three criteria are not pertinent and/or efficient. This leads, in some cases, to some unexplained results. We show the important role, which is mostly neglected in existing method, of frequency content while evaluating the ductile demand. Based on this recognition, we introduce a new method of equivalent linearization based on the transfer function. We use this method in order to explore a numerical experimental plan in which we calculate the equivalent characteristics (frequency and damping) versus the ductile demand for different configuration characterized by (a) the ratio between the frequency of the oscillator and the central frequency of the input signal, (b) the hardening and (c) the behaviour which covers the elastoplastic and damaged ones. We propose two new approaches of the linear equivalent behaviour. The first one, aiming to improve the non-linear static procedure of ATC40, use the secant stiffness and the maximal displacement. This approach consists in an estimation of damping which is different to ATC40. Its pertinence is established by evaluating with accuracy the maximal displacement of the canonical non-linear systems. The second approach consists in restitution of the dynamic of the response of a non-linear oscillator by using the transfer function. The pertinence of this proposition is shown through the criteria of Anderson, especially in terms of transferred motion. In this effect, the linear equivalent behaviour based on the transfer function allows to cope the transferred motion through the non-linear oscillator without performing the non-linear transient calculation. The validation of the linear equivalent behaviour based on the transfer function has been examined on real structures through some experimental tests such as the reinforced concrete wall (SAFE) or piping systems (BARC and EPRI)

Cette thèse porte sur les techniques d'analyse formelle de systèmes hybrides à dynamiques continues non linéaire. Ses contributions portent sur les algorithmes d'atteignabilité et sur les problèmatiques liées à la representation des ensembles atteignables. This thesis deals with formal analysis of hybrid system with non linear continous dynamic. It contributes to the fields of reachability analysis algorithm and the set representation. / In this thesis, we presented our contributions to the formal analysis of dynamical systems. We focused on the problem of efficiently computing an accurate approximation of the reachable sets under nonlinear dynamics given by differential equations. Our aim was also to design scalable methods which can handle large systems. The first contribution of this thesis concerns the dynamic hybridization technique for a large class of nonlinear systems. We focused on the hybridization domain construction such that the linear interpolation realized in this domain ensures a desired error between the original system trajectories and those computed with the approximated system. We propose a construction method which tends to maximize the domain volume which reduce the number of creation of new domains during the analysis. The second research direction that we followed concerns a subclass of nonlinear dynamical systems which are the polynomial systems. Our results for the reachability analysis of these systems are based on the Bernstein expansion properties. We approximate an initial reachability computation (which requires solving polynomial optimization problems) with an accurate over-approximation (which requires solving linear optimization problems). The last theoretical contribution concerns the reachability analysis of linear systems with polyhedral input which often result from approximation of nonlinear systems. We proposed a technique to refine