Search

This work concerns the application of physics-informed neural networks to the modeling and control of complex robotic systems. Achieving this goal requires extending physics-informed neural networks to handle nonconservative effects. These learned models are proposed to combine with model-based controllers originally developed with first-principle models in mind. By combining standard and new techniques, precise control performance can be achieved while proving theoretical stability bounds. These validations include real-world experiments of motion prediction with a soft robot and trajectory tracking with a Franka Emika Panda manipulator., Learning & Autonomous Control

Soft robots are characterized by compliant elements that introduce heightened kinematic complexity compared to their rigid counterparts. Such systems, with infinite degrees of freedom, are inherently underactuated, making precise real-time shape regulation a challenging task. Model-based controllers, utilizing tractable reduced-order modelling methods, have emerged as promising solutions. However, practical implementations of these methods often rely on fully-actuated approximations, overlooking the underactuated nature of these continuum structures. In this study, we aim to experimentally validate model-based controllers that explicitly account for underactuation, surpassing the theoretical feasibility demonstrated in simulation. These controllers incorporate gravity cancellation and compliance compensation using the dynamic model of the robot to achieve superior real-time shape regulation compared to conventional PD/PID controllers. To facilitate this experimental validation, we have built a multi-segment soft robot research platform that includes a passively actuated segment, allowing for the utilization of both actuated and unactuated degrees of freedom in the control feedback loop. Through rigorous experimentation, we provide comprehensive evidence of the efficacy of this class of model-based controllers in controlling unconventionally actuated robotic systems. Consequently, our work bridges the gap between theory and practice, resulting in a practical real-time shape regulation framework that is adaptable to a vast variety of soft robotic systems., Electrical Engineering | Embedded Systems

This research contributes to the field of flexible link parallel robots by showing how a variety of inherent control problems of such robots can be solved. Since the focus lies on promising techniques, well-established approaches are adapted to these robots. The proposed methods are experimentally validated within realistic scenarios. These confirm the need for modern controllers based on flexible multibody models as in most cases they significantly outperform controllers based on classical rigid multibody models., Diese Forschungsarbeit trägt zum Gebiet der parallelen Roboter mit flexiblen Armen bei, indem aufgezeigt wird wie eine Vielzahl von inhärenten Regelungsproblemen solcher Roboter gelöst werden kann. Da der Fokus auf vielversprechenden Techniken liegt, werden etablierte Ansätze an diese Roboter adaptiert. Die vorgebrachten Methoden werden in realistischen Szenarien experimentell validiert. Diese bestätigen den Bedarf an modernen Reglern welche auf flexiblen Mehrkörpermodellen basieren, da sie zumeist Reglern welche auf klassischen starren Mehrkörpermodellen basieren deutlich überlegen sind.

Deformable linear objects (DLOs), such as cables, wires, ropes, and sutures, are important components in various applications in robotics. Although automating DLO manipulation tasks through robot deployment can offer benefits in terms of cost reduction and increased efficiency, it presents difficult challenges. Unlike rigid objects, DLOs can deform and possess high-dimensionalstate space, significantly amplifying the complexity of their dynamics. These inherent characteristics, combined with the absence of distinctive features and the occurrence of occlusion, contribute to the difficulties involved in DLO manipulation tasks. This dissertation focuses on developing novel approaches for two aspects: modeling and tracking DLOs. Both aspects are important in DLO manipulation, yet they remain open research questions. Current analytical physics-based methods for modeling DLO dynamics are either time-consuming or inaccurate and often undifferentiable, which hampers their applications in robot planning and control. Although deep learning methods have shown promise in modeling object dynamics, there is still a gap in learning DLO dynamics in a 3D environment. As for the tracking, many current methods rely on assumptions such as knowing the DLO initial state and segmented DLO point sets, which are rarely fulfilled in real-world scenarios, significantly limiting their practical applicability. This dissertation aims to answer three research questions: How can data-driven models be used for learning DLO dynamics? How can the data-driven models be efficiently trained for real-world DLO manipulation tasks? How can images be used to track the state of DLOs during manipulation in uncontrolled real-world settings? The first contribution of this dissertation is a data-driven model that effectively simulates DLO state transitions. To bridge the current gap in learning full 3D DLO dynamics, a new DLO representation and a recurrent network module are introduced to facilitate better effec

Access to energy is fundamental to economic and technological advancement. Hence, the more the world develops, the greater the demand for energy becomes. Evidently, the production and consumption of energy alone account for more than 80% of global anthropogenic greenhouse gas (GHG) emissions. There is broad scientific consensus that efficiency improvements in energy production and consumption must come first on the path to reducing global GHG emissions. As the largest producer and consumer of energy, the industrial sector faces tremendous challenges due to stringent environmental regulations, intense price-based global competition, rising operating costs and rapidly changing economic conditions. Therefore, increasing energy and resource efficiency while improving throughput and asset reliability is a matter of utmost importance. Satisfying such demanding objectives requires an optimal operation, control and monitoring of plant assets and processes. This is one of the main driving forces behind developing digital solutions, methods, and frameworks that can be integrated with old and new industrial automation platforms. The main focus of this dissertation is to investigate frameworks, process models, soft sensors, control optimization, and diagnostic techniques to improve the operation, control, and monitoring of industrial plants and processes. In this thesis, a generic architecture for control optimization, diagnostics, and decision support system, referred to here as a learning system, is proposed. The research is centred around an investigation of different components of the proposed learning system. Two very different case studies, one representing large-scale assets and another representing a fleet of small-scale assets, are considered to demonstrate the genericness of the proposed system architecture. In this thesis, a very energy-intensive chemical pulping process represents the case study of large-scale assets, and a micro gas turbine (MGT) fleet for distribu, FUDIPO

Reactivity controlled compression ignition (RCCI) is a highly efficient and clean combustion concept, which enables the use of a wide range of renewable fuels. Consequently, this promising dual fuel combustion concept is of great interest for realizing climate neutral future transport. RCCI is very sensitive for operating conditions and requires advanced control strategies to guarantee stable and safe operation. For real-world RCCI implementation, we face control challenges related to transients and varying ambient conditions. Currently, a multivariable air–fuel path controller that can guarantee robust RCCI engine operation is lacking. In this work, we present a RCCI engine controller, which combines static decoupling and a diagonal MIMO feedback controller. For control design, a frequency domain-based approach is presented, which explicitly deals with cylinder-to-cylinder variations using data-driven, cylinder-individual combustion models. This approach enables a systematic trade-off between fast and robust performance and gives clear design criteria for stable operation. The performance of the developed multivariable engine controller is demonstrated on a six-cylinder diesel-E85 RCCI engine. From experimental results, it is concluded that the RCCI engine controller accurately tracks the five desired combustion and air path parameters, simultaneously. For the studied transient cycle, this results in 12.8% reduction in NOx emissions and peak in-cylinder pressure rise rates are reduced by 3.8 bar/deg CA. Compared to open-loop control, the stable and safe operating range is increased from 25 ∘C up to 35 ∘C intake manifold temperature and maximal load range is increased by 14.7% up to BMEP = 14.8 bar

Gas processing facilities, where gas is received and treated, provide motivation and embodiment for the development of systematic control-oriented modeling tools suited to the design of process control solutions based on plant schematics and layouts. The control of these plants involves the interconnection of a number of elements including pipes, compressors, heat exchangers, valves and valve manifolds, and other process units and volumes. The goal of this work is to provide a systematic, scalable and reconfigurable modeling methodology eventually used for process control of such gas-handling facilities at nominal operation. We aim to simplify the control design so that it appeals to generalists without deep expertise in control of fluid flow, using software tools such as Simulink/Matlab and Python. We provide control-oriented component models of standard equipment incl. pipe intersections, compressors, valves and heat exchangers, which serve as modules for entire networks. By exploiting the index-1 property of systems of differential algebraic equations that naturally arise for interconnections our composite LTI state-space models subsume algebraic equations; hence, control-orientation. We also show that for networks conservation of mass is inherited by its components and leads to an integrator in the pressure channel with important implications for control design. The Matlab code provided in the Appendix corroborates the suitability of our approach for software-based control design.

Physically motivated parametric models are the basis of several techniques related to control design. Industrial model-based controller tuning methods include pole placement, symmetric optimum and damping optimum. The challenge is that the resulting model-based controller is satisfactory only if the underlying model is appropriate. Typically, a set of potential models is known a priori, but it is not known, which model should be used. So, the critical question in model-based controller tuning is that of model selection. Existing approaches for model selection are mostly based on maximizing accuracy, but there is no reason why the most accurate model should also be the optimal model for control design. Given the overall aim to design a high-performance controller, in this paper the best model is considered as the one that has the potential to give a model-based controller the highest performance. The proposed method identifies parametric candidate models for control design. Then, a nonparametric model is used to predict the actual performance of the various controllers on the real system. A validation with two industry-like testbeds shows success of the method.

The maritime industry is actively exploring alternative fuels and drive train technology to reduce the emissions of hazardous air pollutants and greenhouse gases. High temperature solid oxide fuel cells (SOFCs) represent a promising technology to generate electric power on ships from a variety of renewable fuels with high efficiencies and no hazardous emissions. However, application in ships is still impeded by a number of challenges, such as low power density and high capital cost. A slow response to load transients is another challenges, which is typically a result of the conservative thermal management strategies used to ensure that excessive thermal stresses in the stack are avoided. Therefore, a reduced order SOFC stack model is developed in this work for model-based control. The model is subsequently verified with a high fidelity model developed in previous work. In addition, a preliminary framework for its use for model predictive SOFC control is provided. The reduced order model and control framework will be used in future work to optimise thermal management of SOFC stacks for improved transient response while respecting physical and operational constraints., Ship Design, Production and Operations, Transport Engineering and Logistics

Modern industrial plants have become increasingly dependent on networked control system architectures in which dedicated sensor‐controller, controller‐actuator, and controller-controller links are replaced by real-time shared (wired or wireless) digital communication networks that operate over specialized industrial networks and protocols. While networked control systems offer a multitude of economic and operational benefits, the increased reliance on shared communication networks comes with a host of fundamental challenges that need to be addressed. For example, challenges such as network resource constraints, data losses, communication delays and real-time scheduling constraints are tied to the inherent limitations on the transmission and processing capabilities of the communication medium, and if left unaddressed can cause operational instabilities or closed-loop performance deterioration. These challenges have motivated a significant and growing body of research work on the analysis and design of networked control systems. An approach that has been proposed to address a number of these challenges is the use of model-based control, however, a central problem that has yet to be addressed is the robustness of the control system to plant-model mismatches that could arise due to things like fouling in heat-exchanger systems or deactivation of catalysts in catalytic reactors.Motivated by these considerations, this dissertation aims to develop an identification-integrated model-based control framework that enhances the performance and security of networked control systems, and address framework implementation issues such as communication resource constraints, lack of full state measurements, distributed systems, communication strategies, nonlinearities, and measurement errors. Finally, the implementation and effectiveness of the developed framework through simulated chemical process examples.

Model based engineering is especially for the development of high performingcontrol systems essentially. By means of suitable simplifications, they help topresent technical relationships and express them mathematically. Thereby,active controllers to influence the system behavior could be developed in anefficient and reliable way. This paper deals with the design of a holistic simulation environment for an ebikewith a wheel hub motor in the rear wheel and a torque and speed sensor inthe bottom bracket. A model-based approach to development using rapidcontrol prototyping is chosen. The model design is chosen similar to the systemdesign of the control system. The interfaces between the main models are alsothe interfaces of the later controller, which makes it easier to implement thesystem afterwards. The engine dynamics has been tested and adjusted on themodel using a driving cycle. A special focus is on the interpretation of thedrivers inputs by the bottom bracket sensor. At the interface between the sensorand the subordinate engine control system, any desired driving condition canbe set for different types of drivers and driving situations by means of differentcharacteristic curves. The scenarios investigated are derived from typical simulations needed duringthe development of e-bike drivetrains. They focus on the interaction of thehybrid system consisting of human driver and engine torque. Especially thesynchronization of torques and the reaction to fast increasing stimuli wereinvestigated. The results show a valid performance of the developed algorithm. The e-motortorque oscillates quickly and the synchronization works fine. Additionally, thealgorithm to smooth the pedaling fluctuations and thereby the torquefluctuations work quite well, whereby a smooth torque is implemented. Nextsteps are the integration of supporting modes and the demand-orientatedcontrol.

A model-based optimal gait is obtained for the 2-D locomotion of a modular snake robot in a duct. Optimality is considered in the sense of traveling as fast as possible or traveling with minimal energy consumption. The novelty of the work lies in the development of a framework to cast the full dynamic behavior, including contact constraints with simple objects, into an optimization problem which allows for gait parameter, control parameter and/or physical parameter optimization. Optimal gait and control parameters are found via a surrogate optimization procedure which reveals optimal locomotion strategies depending on the duct width and optimization criteria. The framework is tested and illustrated with a number of optimizations of 2-D locomotion of a snake robot where either traveling time or energy consumption is minimized.

Learning from data of past tasks can substantially improve the accuracy of mechatronic systems. Often, for fast and safe learning a model of the system is required. The aim of this paper is to develop a model-free approach for fast and safe learning for mechatronic systems. The developed actor-critic iterative learning control (ACILC) framework uses a feedforward parameterization with basis functions. These basis functions encode implicit model knowledge and the actor-critic algorithm learns the feedforward parameters without explicitly using a model. Experimental results on a printer setup demonstrate that the developed ACILC framework is capable of achieving the same feedforward signal as preexisting model-based methods without using explicit model knowledge.

We design, implement and test two control algorithms for autonomously landing a drone on an autonomous boat. The first algorithm uses distributed model predictive control (DMPC), while the second combines a cascade controller with DMPC. The algorithms are implemented on a real drone, while the boat's motion is simulated, and their performance is compared to a centralized model predictive controller. Field experiments are performed, where all algorithms show an ability to land while avoiding violation of the safety constraints. The two distributed algorithms further show the ability to use longer prediction horizons than the centralized model predictive controller, especially in the cascade case, and also demonstrate improved robustness towards breaks in communication., QC 20210710

A characteristic of adaptive behavior is its goal-directed nature. Anability to act in a goal-directed manner is progressively refinedduring development, but this refinement can be impacted by theemergence of psychiatric disorders. Disorders of compulsivity havebeen framed computationally as a deficit in model-based control,and have been linked also to abnormal frontostriatal connectivity.However, the developmental trajectory of model-based control,including an interplay between its maturation and an emergenceof compulsivity, has not been characterized. Availing of a largesample of healthy adolescents (n=569) aged 14 to 24 y, we showbehaviorally that over the course of adolescence there is a within-person increase in model-based control, and this is more pro-nounced in younger participants. Using a bivariate latent changescore model, we provide evidence that the presence of higher com-pulsivity traits is associated with an atypical profile of this develop-mental maturation in model-based control. Resting-state fMRI datafrom a subset of the behaviorally assessed subjects (n=230)revealed that compulsivity is associated with a less pronouncedchange of within-subject developmental remodeling of functionalconnectivity, specifically between the striatum and a frontoparietalnetwork. Thus, in an otherwise clinically healthy population sample,in early development, individual differences in compulsivity arelinked to the developmental trajectory of model-based controland a remodeling of frontostriatal connectivity.

Driven by intense competition, increasing operational cost and strict environmental regulations, the modern process and energy industry needs to find the best possible way to adapt to maintain profitability. Optimization of control and operation of the industrial systems is essential to satisfy the contradicting objectives of improving product quality and process efficiency while reducing production cost and plant downtime. Use of optimization not only improves the control and monitoring of assets but also offers better coordination among different assets. Thus, it can lead to considerable savings in energy and resource consumption, and consequently offer a reduction in operational costs, by offering better control, diagnostics and decision support. This is one of the main driving forces behind developing new methods, tools and frameworks that can be integrated with the existing industrial automation platforms to benefit from optimal control and operation. The main focus of this dissertation is the use of different process models, soft sensors and optimization techniques to improve the control, diagnostics and decision support for the process and energy industry. A generic architecture for an optimal control, diagnostics and decision support system, referred to here as a learning system, is proposed. The research is centred around an investigation of different components of the proposed learning system. Two very different case studies within the energy-intensive pulp and paper industry and the promising micro-combined heat and power (CHP) industry are selected to demonstrate the learning system. One of the main challenges in this research arises from the marked differences between the case studies in terms of size, functions, quantity and structure of the existing automation systems. Typically, only a few pulp digesters are found in a Kraft pulping mill, but there may be hundreds of units in a micro-CHP fleet. The main argument behind the selection of these two case, FUDIPO – FUture DIrections for Process industry Optimization

Symbolic models are becoming more and more popular in the research community working on hybrid systems because they provide a systematic approach to enforce logic specifications on purely continuous or hybrid systems while fulfilling the constraints at the hardware/software implementation level. This paper contributes to this research line and proposes symbolic models approximating possibly unstable time-delay systems with quantized measurements of the outputs. An application of the proposed results to the glucose control problems for the Artificial Pancreas is discussed in the paper.

In this paper we deal with the problem of tracking a desired plasma glucose evolution by means of intra-venous insulin administration, for Type 2 diabetic patients exhibiting basal hyperglycemia. A nonlinear time-delay model is used to describe the glucose-insulin regulatory system, and a modelbased approach is exploited in order to design a global sampleddata control law for such system. Sontag's universal formula is designed to obtain a steepest descent feedback induced by a suitable control Lyapunov-Krasovskii functional. Such a feedback is a stabilizer in the sample-and-hold sense. Furthermore, the input-to-state stability redesign method is used in order to attenuate the effects of bounded actuation disturbances and observation errors, which can appear for uncertainties in the instruments. The proposed control law depends on sampled glucose and insulin measurements. Theoretical results are validated through simulations.

System identification is a common tool for estimating (linear) plant models as a basis for model-based predictive control and optimization. The current challenges in process industry, however, ask for data-driven modelling techniques that go beyond the single unit/plant models. While optimization and control problems become more and more structured in the form of decentralized and/or distributed solutions, the related modelling problems will need to address structured and interconnected systems. An introduction will be given to the current state of the art and related developments in the identification of linear dynamic networks. Starting from classical prediction error methods for open-loop and closed-loop systems, several consequences for the handling of network situations will be presented and new research questions will be highlighted.

In this work, we consider the problem of controlling a two-stage cooling mixed suspension mixed product removal (MSMPR) crystallizer. For this process, the temperature in the first crystallizer is manipulated for controlling the average crystal dimension (d43), while it is desired to maintain the temperature of the second crystallizer at the minimum allowed value to guarantee maximum yield. Due to system nonlinearities and process delays, the performance of traditional PID controllers are poor. A control scheme is proposed to improve the closed loop performance and achieve desired control objectives. The control scheme is based on the coupling of a PI controller and a model-based nonlinear prediction block that serves as delay and disturbance compensator. The proposed scheme has been tested on the system in case of disturbances in the feed concentration, with and without measurement errors. It is observed that the proposed scheme outperforms the PI controller by reducing the output response settling time and overshoot.

One of the main goals of laminar flow control is to reduce skin-friction drag by delaying the onset of laminar to turbulent transition. Over unswept wings, the leading cause of this is the growth of flow instabilities called Tollmien-Schlichting(TS) waves. This thesis aims to test the feasibility and performance of a designed Linear Quadratic Gaussian (LQG) controller in attenuating TS waves, in an experimental setting. A modified flat-plate test setup with an adverse pressure gradient was used to generate and characterize these waves, while control was performed using a Dielectric-Barrier Discharge (DBD) plasma actuator. Embedded microphones were used to measure the pressure fluctuations within the boundary layer, while Particle Image Velocimetery (PIV) was used to quantify the flow. The performance of the controller was investigated at its nominal and off-design conditions (robustness), as well as its comparison to open-loop continuous forcing. For the nominal design conditions, a reduction in the spectral energy of the peak frequencies of four orders of magnitude are observable for the closed-loop case, two orders more than the open-loop case. The root-mean-square (RMS) of the downstream signals showed an overall maximum additional reduction of 55% of the pressure fluctuations, compared to the open-loop case. The controller was more robust at lower peak-to-peak voltages, with an overall 30-60% reduction in RMS of the pressure fluctuations, compared to open-loop forcing .This study demonstrates the feasibility of using LQG model-based techniques as a viable flow-control strategy in damping flow-instabilities, and is an option worth further investigation., Aerospace Engineering

Model-based controllers with adaptive design variables are often used to control an object with time-dependent characteristics. However, the controller's performance is influenced by many factors such as modeling accuracy and fluctuations in the object's characteristics. One method to overcome these negative factors is to tune model-based controllers. Herein we propose an online tuning method to maintain control performance for an object that exhibits time-dependent variations. The proposed method employs the poles of the controller as design variables because the poles significantly impact performance. Specifically, we use the simultaneous perturbation stochastic approximation (SPSA) to optimize a model-based controller with multiple design variables. Moreover, a vibration control experiment of an object with time-dependent characteristics as the temperature is varied demonstrates that the proposed method allows adaptive control and stably maintains the closed-loop characteristics.

An advanced model-based control method for the integrated fuel processing and a fuel cell system consisting of ethanol reforming, hydrogen purification, and a proton exchange membrane fuel cell is presented. For process identification, a physical model of the process chain was constructed. Subsequently, the simulated process was approximated with data-driven control models. Based on these control models, a hierarchical control framework consisting of model predictive controller and a global optimization algorithm was introduced. The performance of the new control method was evaluated with simulations. Results indicate that the new optimization concept enables resource efficient and fast control of the studied energy conversion process. Fast and efficient fuel cell process could then provide sustainable power source for autonomous and mobile applications in the future.

Model-based controllers with adaptive design variables are often used to control an object with time-dependent characteristics. However, the controller's performance is influenced by many factors such as modeling accuracy and fluctuations in the object's characteristics. One method to overcome these negative factors is to tune model-based controllers. Herein we propose an online tuning method to maintain control performance for an object that exhibits time-dependent variations. The proposed method employs the poles of the controller as design variables because the poles significantly impact performance. Specifically, we use the simultaneous perturbation stochastic approximation (SPSA) to optimize a model-based controller with multiple design variables. Moreover, a vibration control experiment of an object with time-dependent characteristics as the temperature is varied demonstrates that the proposed method allows adaptive control and stably maintains the closed-loop characteristics.

The Ampelmann system offers a safe and reliable solution for offshore access. As safety is a key factor, an extensive safety and warning management system is employed which accommodates for different fault types that may occur. Accommodation of sensor failure in the Ampelmann system is currently done through switching to a redundant component. However, there are several faults that are left undetected in the system. Detection only occurs when the faults exceed a critical threshold. Exceeding this threshold immediately results in shut-down of the system as safety is no longer guaranteed. For sensor equipment critical to the motion compensation, this leads to a code black in the system. The occurrences of code blacks should be limited where possible due to the fact that these lead to downtime. A critical sensor for the motion control is the position transducer in the hydraulic cylinders. The measured lengths are used for feedback purposes in the control system. The position transducer is redundant in each cylinder. The redundant sensor is mainly utilized for checking the main sensor. However, when the measurements from both sensors deviate too much from one another the system will shut down. Therefore, in this thesis, the possibility of a model-based fault detection method for the position transducer in the hydraulic cylinder is explored. Firstly, an accurate model of the hydraulic cylinder is derived and identified. Then, the model is combined with an observer to generate accurate estimates of the cylinder lengths. Furthermore, the estimates are compared with the actual measured cylinder lengths from the position transducer to generate residuals. Finally, the residuals are evaluated in order to make a decision about the health of the sensor. Three different fault types have been defined, which are expected to cause sensor degradation/failure. For each fault type, the residuals are evaluated. Prior to this a threshold has been defined based, Mechanical Engineering | Systems and Control

This thesis report proposes a framework to implement Navigation, Guidance and Control (GNC) systems, that enable point-to-point autonomy for displacement vessels. A model-based control approach is chosen as the basis of the GNC systems. The resulting algorithms are implemented for verification in a 1:25 scale model of a Azimuth Stern Drive (ASD) 3111 Damen tug named "Damen Autonomous Ship", aka DASh. First, a compact maneuvering model that captures relevant dynamics of displacement vessel is formulated and identified using system identification. Secondly,the guidance system is automated such that it connects an initial state to a goal state with a collision free path that satisfies all input and differential constraints of the vessel model. To this end, the kinodynamic Rapidly-exploring Random Tree (RRT) algorithm is extended to use a maneuver automaton and optimal motion primitives in its steering function. A learned cost-to-go distance metric for the state space is formulatedto efficiently calculate distance between states, which is used to search for nearest neighbors in the kino- dynamic RRT algorithm. The performance of the planner using the learned cost-to-go distance metric is compared to a minimal curve length distance metric based on Dubins Curves and the commonly used straight-line Euclidean distance metric. It is shown that the learned cost-to-go and the minimal curve length distance metric result in paths of similar performance while the Euclidean metric performs severely worse. Lastly, the navigation and control systems are implemented on DASh. Due to disturbances present in real world environments, the paths must be tracked using feedback control. State estimation for navigation, based on position and heading measurements is performed by implementing an observer using a Extended Kalman Filter (EKF). Non-linear model predictive control (NMPC) in combination with thrust allocation is used to control

The method of Lagrangian particle tracking (LPT) is extended to autonomous underwater vehicles (AUVs) that are modeled as controlled particles. Controlled LPT (CLPT) evaluates the performance of ocean models used for the navigation of AUVs by computing the differences between predicted vehicle trajectories and actual vehicle trajectories. Such difference, measured by the controlled Lagrangian prediction error (CLPE), demonstrates growth rate that is influenced by the accuracy of the ocean model and the strength of the feedback control laws used for vehicle navigation. Despite the limited accuracy and resolution of the ocean model, the error growth can be bounded by feedback control when localization service is available, which is a unique property of CLPT. Theoretical relationship among CLPE growth rate, quality of ocean models, and feedback control are established for two control strategies: a transect-following controller and a station-keeping controller that are often used by AUVs. Upper bounds for error growth are derived and verified by both simulation and experimental data collected during the operations of underwater gliders in coastal ocean. © 2017 IEEE.

[Resumen] La estimación de los estados no observables de un proceso es importante cuando se usan técnicas de control que suponen dichos valores conocidos apriori. Los controladores basados en el espacio de estados presentan un buen comportamiento y eficiencia, aún cuando la dinámica del proceso es no lineal. Uno de los procesos con una dinámica no lineal y en la que no todos los estados son observables es el caso de las plantas solares cilindro- parabólicas. En este trabajo se presenta un observador, basado en un sistema de inferencia borroso, para la estimación de los perfiles de temperatura de los lazos que componen el campo solar., [Abstract] The estimation of the unobservable states of a process is important when using control techniques that assume, a priori, these values are known. The controllers based on the state space are very efficient and even more when the dynamics of the process are strongly non-linear. One of the processes with very non-linear dynamics, and in which not all the states are observable, is the case of parabolic trough solar plants. In this work an observer based on Fuzzy Logic is presented to estimate the temperature profiles of the loops that form the solar field

Real-time optimization of systems governed by partial differential equations (PDEs) presents significant computational challenges to nonlinear model predictive control (NMPC). The large-scale nature of PDEs often limits the use of standard nested black-box optimizers that require repeated forward simulations and expensive gradient computations. Hence, to ensure online solutions at relevant time-scales, large-scale NMPC algorithms typically require powerful, customized PDE-constrained optimization solvers. To this end, this paper proposes a new Lagrange-Newton-Krylov (LNK) method that targets the class of time-dependent nonlinear diffusion-reaction systems arising from chemical processes. The LNK solver combines a high-order spectral Petrov-Galerkin (SPG) method with a new, parallel preconditioner tailored for the large-scale saddle-point systems that form subproblems of Sequential Quadratic Programming (SQP) methods. To establish proof-of-concept, a case study uses a simple parallel MATLAB implementation of the preconditioner with 10 cores. As a step towards real-time control, the results demonstrate that large-scale diffusion-reaction optimization problems with more than 106 unknowns can be solved efficiently in less than a minute.

[EN] Rehabilitation is a hazardous task for a mechanical system, since the device has to interact with the human extremities without the hands-on experience the physiotherapist acquires over time. A gap needs to be filled in terms of designing effective controllers for this type of devices. In this respect, the paper describes the design of a model-based control for an electromechanical lower-limb rehabilitation system based on a parallel kinematic mechanism. A controller-observer was designed for estimating joint velocities, which are then used in a hybrid position/force control scheme. The model parameters are identified by customising an approach based on identifying only the relevant system dynamics parameters. Findings obtained through simulations show evidence of improvement in tracking performance compared with those where the velocity was estimated by numerical differentiation. The controller is also implemented in an actual electromechanical system for lower-limb rehabilitation tasks. Findings based on rehabilitation tasks confirm the findings from simulations.

This paper presents a novel approach to deal with the regulation of the dc-link capacitor voltages and ac-side currents in a grid-connected five-level diode-clamped converter. Due to the controllability problems of this topology, guaranteeing a solid current control and, mainly, a correct dc-link voltage sharing, represents a complex technical challenge. With the purpose of coping with it, an averaged model that describes the system dynamics at both sides of the converter is presented, assuming that a modulation strategy is integrated in the system to generate the switching sequence. In order to derive the proposed model, no restriction concerning the use of only the three nearest vectors to the desired voltage reference is taking into account. Then, several changes of variables are carried out in the model equations to obtain control input decoupling for control purposes, while reducing the complexity of the model as well. Finally, the voltage and current controllers are designed separately using different control inputs in a straightforward way. Neither auxiliary hardware nor complicated mathematical calculations are required to achieve the control objectives. The effectiveness and good performance of the system under the proposed control approach is validated by simulation results, suggesting that the five-level diode-clamped converter can be a solid solution as an interfacing system connected to the utility grid for, e.g., industrial drives or renewable energy applications.

Representation learning is a central topic in the field of deep learning. It aims at extracting useful state representations directly from raw data. In deep learning, state representations are usually used for classification or inferences. For example, image embedding that provides similarity metrics can be used for face recognition. Recent success on deep learning has stimulated the interest of applying state representation learning to control. This problem is very different from deep learning in computer visions and language modelling. Control tasks usually require pose information about the task relevant object in the scene. Applying state representation learning to control requires such features to be embedded. This is a difficult problem since there is generally no explicit supervision for extracting pose information from the data. A problem of state representation learning for control is that raw data can contain much irrelevant information for the control task. For example, colors, textures and shapes of task-relevant objects are not as important as poses of objects. This suggests that the appearance and pose information should be disentangled in learned representations and the pose representations should be used for the control task. Furthermore, we usually need a system dynamic model in order to perform model-based optimal control. The prediction ability of the model is crucial for planning since it has to evaluate future trajectories and optimizes the actions. Thus we need to learn a decent dynamic model on extracted representations. To address these problems, we first propose to use Variational Auto-encoders to disentangle the pose and appearance representations. The benefit of preserving both appearance and pose representations is that we can predict pose representations conditioned on actions. From appearance and predicted pose representations, we can reconstruct and predict future image frames. For this purpose, we introduce the Long

Nowadays in large-scale process industry, high purity of products is not only desired but crucial. Products must meet high-purity standards to conform with market and customers’ requirements. Two chemical processes, reaction and distillation, are essential to achieve these objectives. To overcome certain disadvantages and/or limitations that each operation unit possess, these two processes are sometimes combined in one. The combination of a separation zone with a reaction zone in the reactive distillation column, leads to complex interactions between vapor-liquid equilibrium, mass transfer rates, diffusion and chemical kinetics, which are accurately described by rigorous mathematical models. Nevertheless, this kind of models are inconvenient when it comes to parameter estimation and model-based control implementation; controllers and estimators require simpler models to perform reliably and efficiently. This situation poses a trade-off problem between model accuracy and control efficiency and implementability. The extents of reaction and flows is a mathematical framework that has been attracting a lot of attention in recent years, especially in processes with chemical reaction. This approach introduces a linear transformation to decompose the system dynamics into its reaction and flow spaces to obtain a decoupled quasi-linear representation of the system’s governing dynamics. The nonlinear model is brought to an LPV description by means of this transformation. Due to the decoupling effect on the global dynamics, the system is simplified in representation and no reduction of the state-space dimension or linearization needs to be done. The extent transformation framework is extended to the case of a reactive batch distillation column to build an LPV model for control. This model is used to develop an MPC. To test the efficiency of the LPV-based scheme, the MPC is compared against a nonlinear MPC., IMPROVISE

Aplicat embargament des del moment de la defensa fins al 5 de juliol de 2019., In recent years, the PEM fuel cell technology has been incorporated to the R&D plans of many key companies in the automotive, stationary power and portable electronics sectors. However, despite current developments, the technology is not mature enough to be significantly introduced into the energy market. Performance, durability and cost are the key challenges. The performance and durability of PEM fue! cells significantly depend on variations in the concentrations of hydrogen and oxygen in the gas channels, water activity in the catalyst layers and other backing layers, water content in the polymer electrolyte membrane, as well as temperature, among other variables. Such variables exhibit intemal spatial dependence in the direction of the fuel and air streams of the anode and cathode. Highly non-uniform spatial distributions in PEM fuel cells result in local over-heating, cell flooding, accelerated ageing, and lower power output than expected. Despite the importance of spatial variations of certain variables in PEM fuel cells, not many works available in the literature target the control of spatial profiles. Most control-oriented designs use lumped-parameter models because of their simplicity and convenience for controller performance. In contrast, this Doctoral Thesis targets the distributed parameter modelling and control of PEM fuel cells. In the modelling part, the research addresses the detailed development of a non-linear distributed parameter model of a single PEM fuel cell, which incorporates the effects of spatial variations of variables that are relevant to its proper performance. The model is first used to analyse important cell intemal spatial profiles, and it is later simplified in arder to decrease its computational complexity and make it suitable for control purposes. In this task, two different model order reduction techniques are applied and compared. The purpose of the control part is to tackle water management and supply of reactants, which are t, A pesar de los avances actuales, la tecnología de celdas de hidrógeno tipo PEM no está suficientemente preparada para ser ampliamente introducida en el mercado energético. Rendimiento, durabilidad y costo son los mayores retos. El rendimiento y la durabilidad de las celdas dependen significativamente de las variaciones en las concentraciones de hidrógeno y oxígeno en los canales de alimentación de gases, la humedad relativa en las capas catalizadoras, el contenido de agua de la membrana polimérica, así como la temperatura, entre otras variables. Dichas variables presentan dependencia espacial interna en la dirección del flujo de gases del ánodo y del cátodo. Distribuciones espaciales altamente no uniformes en algunas variables de la celda resultan en sobrecalentamiento local, inundación, degradación acelerada y menor potencia de la requerida. Muy pocos trabajos disponibles en la literatura se ocupan del control de perfiles espaciales. La mayoría de los diseños orientados a control usan modelos de parámetros concentrados que ignoran la dependencia espacial de variables internas de la celda, debido a la complejidad que añaden al funcionamiento de controladores. En contraste, esta Tesis Doctoral trata la modelización y control de parámetros distribuidos en las celdas de hidrógeno tipo PEM. En la parte de modelización, esta tesis presenta el desarrollo detallado de un modelo no lineal de parámetros distribuidos para una sola celda, el cual incorpora las variaciones espaciales de todas las variables que son relevantes para su correcto funcionamiento. El modelo se usa primero para analizar importantes perfiles espaciales internos, y luego se simplifica para reducir su complejidad computacional y adecuarlo a propósitos de control. En esta tarea se usan y se comparan dos técnicas de reducción de orden de modelos. El propósito de la parte de control es abordar la gestión de agua y el suministro de reactantes, que son dos grandes retos en el funcionamiento de las celdas con i, Postprint (published version)

[EN]: In recent years, the PEM fuel cell technology has been incorporated to the R&D plans of many key companies in the automotive, stationary power and portable electronics sectors. However, despite current developments, the technology is not mature enough to be significantly introduced into the energy market. Performance, durability and cost are the key challenges. The performance and durability of PEM fuel cells significantly depend on variations in the concentrations of hydrogen and oxygen in the gas channels, water activity in the catalyst layers and other backing layers, water content in the polymer electrolyte membrane, as well as temperature, among other variables. Such variables exhibit internal spatial dependence in the direction of the fuel and air streams of the anode and cathode. Highly non-uniform spatial distributions in PEM fuel cells result in local over-heating, cell flooding, accelerated ageing, and lower power output than expected. Despite the importance of spatial variations of certain variables in PEM fuel cells, not many works available in the literature target the control of spatial profiles. Most controloriented designs use lumped-parameter models because of their simplicity and convenience for controller performance. In contrast, this Doctoral Thesis targets the distributed parameter modelling and control of PEM fuel cells. In the modelling part, the research addresses the detailed development of a non-linear distributed parameter model of a single PEM fuel cell, which incorporates the effects of spatial variations of variables that are relevant to its proper performance. The model is first used to analyse important cell internal spatial profiles, and it is later simplified in order to decrease its computational complexity and make it suitable for control purposes. In this task, two different model order reduction techniques are applied and compared. The purpose of the control part is to tackle water management and supply of reactants, whic, [ES]: A pesar de los avances actuales, la tecnología de celdas de hidrógeno tipo PEM no está suficientemente preparada para ser ampliamente introducida en el mercado energético. Rendimiento, durabilidad y costo son los mayores retos. El rendimiento y la durabilidad de las celdas dependen significativamente de las variaciones en las concentraciones de hidrógeno y oxígeno en los canales de alimentación de gases, la humedad relativa en las capas catalizadoras, el contenido de agua de la membrana polimérica, así como la temperatura, entre otras variables. Dichas variables presentan dependencia espacial interna en la dirección del flujo de gases del ánodo y del cátodo. Distribuciones espaciales altamente no uniformes en algunas variables de la celda resultan en sobrecalentamiento local, inundación, degradación acelerada y menor potencia de la requerida. Muy pocos trabajos disponibles en la literatura se ocupan del control de perfiles espaciales. La mayoría de los diseños orientados a control usan modelos de parámetros concentrados que ignoran la dependencia espacial de variables internas de la celda, debido a la complejidad que añaden al funcionamiento de controladores. En contraste, esta Tesis Doctoral trata la modelización y control de parámetros distribuidos en las celdas de hidrógeno tipo PEM. En la parte de modelización, esta tesis presenta el desarrollo detallado de un modelo no lineal de parámetros distribuidos para una sola celda, el cual incorpora las variaciones espaciales de todas las variables que son relevantes para su correcto funcionamiento. El modelo se usa primero para analizar importantes perfiles espaciales internos, y luego se simplifica para reducir su complejidad computacional y adecuarlo a propósitos de control. En esta tarea se usan y se comparan dos técnicas de reducción de orden de modelos. El propósito de la parte de control es abordar la gestión de agua y el suministro de reactantes, que son dos grandes retos en el funcionamiento de las celdas

A novel model-based control strategy has been developed for filamentous fungal fed-batch fermentation processes. The system of interest is a pilot scale (550 L) filamentous fungus process operating at Novozymes A/S. In such processes, it is desirable to maximize the total product achieved in a batch in a defined process time. In order to achieve this goal, it is important to maximize both the product concentration, and also the total final mass in the fed-batch system. To this end, we describe the development of a control strategy which aims to achieve maximum tank fill, while avoiding oxygen limited conditions. This requires a two stage approach: (i) calculation of the tank start fill; and (ii) on-line control in order to maximize fill subject to oxygen transfer limitations. First, a mechanistic model was applied off-line in order to determine the appropriate start fill for processes with four different sets of process operating conditions for the stirrer speed, headspace pressure, and aeration rate. The start fills were tested with eight pilot scale experiments using a reference process operation. An on-line control strategy was then developed, utilizing the mechanistic model which is recursively updated using on-line measurements. The model was applied in order to predict the current system states, including the biomass concentration, and to simulate the expected future trajectory of the system until a specified end time. In this way, the desired feed rate is updated along the progress of the batch taking into account the oxygen mass transfer conditions and the expected future trajectory of the mass. The final results show that the target fill was achieved to within 5% under the maximum fill when tested using eight pilot scale batches, and over filling was avoided. The results were reproducible, unlike the reference experiments which show over 10% variation in the final tank fill, and this also includes over filling. The variance of the final tank fill is reduced

Simulation studies of oil field water flooding have demonstrated a significant potential of optimal control technology to improve industrial practices. However, real-life applications are challenged by unknown geological factors that make reservoir models highly uncertain. To minimize the associated financial risks, the oil literature has used ensemble-based methods to manipulate the net present value (NPV) distribution by optimizing sample estimated risk measures. In general, such methods successfully reduce overall risk. However, as this paper demonstrates, ensemble-based control strategies may result in individual profit outcomes that perform worse than real-life dominating strategies. This poses significant financial risks to oil companies whose main concern is to avoid unacceptable low profits. To remedy this, this paper proposes offset risk mimimization. Unlike existing methodology, the offset method uses the NPV offset distribution to minimize risk relative to a competing reference strategy. Open-loop simulations of a 3D two-phase synthetic reservoir demonstrate the potential of offset risk minimization to significantly improve the worst case profit offset relative to real-life best practices. The results suggest that it may be more relevant to consider the NPV offset distribution than the NPV distribution when minimizing risk in production optimization.

[ES] El objetivo principal del presente trabajo ha sido el de elaborar una estrategia de control para aplicar a un proceso de filtración en un MBR anaeróbico sumergido (SAnMBR). Éste, tiene como finalidad controlar el FR del sistema de filtración mediante la optimización del caudal de gas a recircular. Para llevarlo a cabo, se ha basado en el modelo de tres resistencias en serie i contemplando distintas condiciones de operación para encontrar aquellas que sean económicamente más óptimas. Las condiciones de operación con las que se ha trabajado han sido la concentración de sólidos en el licor mezcla al tanque de membranas y tiempo de filtrado. Para la simulación y validación de la estrategia de control se ha utilizado la modelización de una planta piloto SAnMBR realizada por el grupo CalAgua y equipada con membranas de ultrafilación de fibra hueca de escala industrial. Se han contemplado los costes de funcionamiento, de la vida útil de las membranas y del coste de los lavados químicos como factores para encontrar aquellas condiciones que sean económicamente más favorables., [CA] L’objectiu principal del present treball ha estat l’elaboració d’una estratègia de control per a aplicar al procés de filtració en un MBR anaeròbic submergit (SAnMBR). Aquesta, te com a finalitat controlar el FR del sistema de filtració mitjançant l’optimització del cabal de gas a recircular. Per a fer-ho, s’ha basat en el model de resistències en sèrie i contemplant diferents condicions d’operació per a trobar-ne aquelles que siguin econòmicament més òptimes. Les condicions d’operació amb les que s’ha treballat han estat la concentració de sòlids en el licor mescla al tanc de membranes i la durada de les fases de filtració. Per a la simulació i validació de l’estratègia de control s’ha fet ús de la modelització d’una planta pilot SAnMBR elaborada pel grup CalAgua i equipada amb membranes d’ultrafiltració de fibra buida d’escala industrial. S’han contemplats els costs de funcionament, de la vida útil de les membranes i del cost dels rentats químics com a factors per a trobar aquelles condicions que siguin econòmicament més favorables., [EN] The main aim of this final master project has been to develop a control strategy for a filtration process in a submerged anaerobic MBR (SAnMBR). The aim is to control the FR of the filtration system by optimizing biogas recycling flow (BRF). To perform this task, it has been based on three resistance-in-series model and worked under different operation conditions to find those that are economically optimal. The operating conditions in which we have worked with have been the solid concentration in the mixed liquor tank and the filtration time. For the control strategy simulation and validation it has been used a SAnMBR pilot plant modeled by CalAgua group and equipped with industrial-scale hollow-fiber ultrafiltration membranes. The team has contemplated operating costs, membranes service-life and chemical-washes costs to find those conditions which are economically more favorable.

Transition delay in boundary-layer flows is achieved via reactive control of flow instabilities, i.e. Tollmien-Schlichting (TS) waves. Adaptive and model-based control techniques are investigated by means of direct numerical simulations (DNS) and experiments. The action of actuators localised in the wall region is prescribed based on localised measurement of the disturbance field; in particular, plasma actuators and surface hot-wire sensors are considered. Performances and limitations of this control approach are evaluated both for two-dimensional (2D) and three-dimensional (3D) disturbance scenarios. The focus is on the robustness properties of the investigated control techniques; it is highlighted that static model-based control, such as the linear-quadratic- Gaussian (LQG) regulator, is very sensitive to model-inaccuracies. The reason for this behaviour is found in the feed-forward nature of the adopted sensor/actuator scheme; hence, a second, downstream sensor is introduced and actively used to recover robustness via an adaptive filtered-x least-mean-squares (fxLMS) algorithm. Furthermore, the model of the flow required by the control algorithm is reduced to a time delay. This technique, called delayed-x least-mean-squares (dxLMS) algorithm, allows taking a step towards a self-tuning controller; by introducing a third sensor it is possible to compute on-line the suitable time-delay model with no previous knowledge of the controlled system. This self-tuning approach is successfully tested by in-flight experiments on a motor-glider. Lastly, the transition delay capabilities of the investigated control con- figuration are confirmed in a complex disturbance environment. The flow is perturbed with random localised disturbances inside the boundary layer and the laminar-to-turbulence transition is delayed via a multi-input-multi-output (MIMO) version of the fxLMS algorithm. A positive theoretical net-energy- saving is observed for disturbance amplitudes up to 2% of the, I den här avhandlingen har reglertekniska metoder tillämpats för att försena omslaget från ett laminärt till ett turbulent gränsskikt genom att dämpa tillväxten av små instabiliteter, så kallade Tollmien-Schlichting vågor. Adaptiva och modellbaserade metoder för reglering av strömning har undersökts med hjälp av numeriska beräkningar av Navier-Stokes ekvationer, vindtunnelexperiment och även genom direkt tillämpning på flygplan. Plasmaaktuatorer och varmtrådsgivare vidhäftade på ytan av plattan eller vingen har använts i experimenten och modellerats i beräkningarna. Prestanda och begränsningar av den valda kontrollstrategin har utvärderats för både tvådimensionella och tredimensionella gränsskiktsinstabiliteter. Fokus har varit på metodernas robusthet, där vi visar att statiska metoder som linjär-kvadratiska regulatorer (LQG) är mycket känsliga för avvikelser från den nominella modellen. Detta beror främst på att regulatorer agerar i förkompenseringsläge (”feed-foward”) på grund av strömningens karaktär och placeringen av givare och aktuatorer. För att minska känsligheten mot avvikelser och därmed öka robustheten har en givare införts nedströms och en adaptiv fXLMS algoritm (filtered-x least-mean-squares) har tillämpats. Vidare har modelleringen av fXLMS-algoritmen förenklats genom att ersätta överföringsfunktionen mellan aktuatorer och givare med en lämplig tidsfördröjning. Denna metod som kallas för dxLMS (delayed-x least-mean-squares) kräver att ytterligare en givare införs långt uppströms för att kunna uppskatta hastigheten på de propagerande instabilitetsvågorna. Denna teknik har tillämpats framgångsrikt för reglering av gränsskiktet på vingen av ett segelflygplan. Slutligen har de reglertekniska metoderna testas för komplexa slumpmässiga tredimensionella störningar som genererats uppströms lokalt i gränsskiktet. Vi visar att en signifikant försening av laminärt-turbulentomslag äger rum med hjälp av en fXLMS algoritm. En analys av energibudge

Fuel economy standards for cars and other vehicles are growing increasingly stringent, thus motivating automakers to find ways to improve fuel efficiency. One popular strategy is to turbocharge a downsized (smaller displacement) engine, which can be more fuel efficient than a naturally aspirated engine delivering the same power output. However, turbocharged engines can be sluggish to respond to torque requests, which drivers often find undesirable. Unfortunately, improving torque responsiveness results in reduced fuel efficiency, and vice versa.This dissertation explores two model-based control strategies to manage this tradeoff. The first strategy is a decentralized controller, in which the throttle and wastegate are controlled in separate loops. The throttle loop uses feedback linearization with supplemental PI control to obtain good torque tracking. The wastegate is opened or closed, based on a preview of the reference torque, to switch between fuel-optimal and torque-optimal modes. The second strategy is a multi-objective optimization scheme to obtain good fuel efficiency and fast torque response by controlling the throttle and wastegate simultaneously. Simulation results show promising performance from both strategies.Additionally, the models used in these control methods are described in detail. A high-fidelity engine simulator in Simscape is used for controller validation. This simulator is too complex for controller design, so a simpler 4-state model is constructed. This model works well in continuous time, but the optimization-based control method requires a discrete-time model. Unfortunately, discretizing the 4-state model results in chattering due to numerical stiffness. This numerical stiffness is analyzed, and a solution is proposed to represent the throttle pressure ratio as a static map. This results in a 3-state model that is easily discretized.

Model predictive control (MPC) is an advanced control technique that offers an elegant framework to solve a wide range of control problems (regulation, tracking, supervision, etc) and handle constraints on the plant. The control objectives and constraints are usually formulated as an optimization problem that the MPC controller has to solve (either offline or online) to return the control command for the plant. This master thesis proposes a novel primal-dual interior-point (PDIP) method for solving quadratic programming problems with linear inequality constraints that typically arise from MPC applications. Convergence of PDIP is studied both in primal and dual framework. We show that the solver converges quadratically to a suboptimal solution of the MPC problem. PDIP solvers rely on two phases: the damped and the pure Newton phases. Compared to state-of-the-art PDIP method, this new solver replaces the initial (linearly convergent) damped Newton phase (usually used to compute a medium-accuracy solution) with a dual solver based on Nesterov's fast gradient scheme (DFG) that converges super-linearly to a medium-accuracy solution. The switching strategy to the pure Newton phase, compared to the state of the art, is computed in the dual space to exploit the dual information provided by the DFG in the first phase. Removing the damped Newton phase has the additional advantage that this solver saves the computational effort required by backtracking line search. The effectiveness of the proposed solver is demonstrated by simulating it on a 2-dimensional discrete-time unstable system., Mechanical, Maritime and Materials Engineering, Delft Center for Systems and Control (DCSC)

This paper presents an image-based controller to perform the guidance of a free-floating robot manipulator. The manipulator has an eye-in-hand camera system, and is attached to a base satellite. The base is completely free and floating in space with no attitude control, and thus, freely reacting to the movements of the robot manipulator attached to it. The proposed image-based approach uses the system's kinematics and dynamics model, not only to achieve a desired location with respect to an observed object in space, but also to follow a desired trajectory with respect to the object. To do this, the paper presents an optimal control approach to guiding the free-floating satellite-mounted robot, using visual information and considering the optimization of the motor commands with respect to a specified metric along with chaos compensation. The proposed controller is applied to the visual control of a four-degree-of-freedom robot manipulator in different scenarios.

Plug-fow crystallization (PFC) is a promising continuous pharmaceutical crystal- lization process but is prone to fouling due to uncontrolled crystallization on the reactor surface (encrustation). This results in operational issues such as (1) fow blockage, (2) increased thermal resistance, and (3) reduced supersaturation, and in turn leads to limited continuous operation and reduced crystal quality and yield. In this work, we introduce a model-based anti-fouling control (AFC) via spatial and temporal heating and cooling cycle.

Cloud computing is an application hosting model providing the illusion of infinite computing power. However, even the largest datacenters have finite computing capacity, thus cloud infrastructures have experienced overload due to overbooking or transient failures. The topic of this paper is the comparison of different control strategies to mitigate overload for datacenters, that assume that the running cloud applications are cooperative and help the infrastructure in recovering from critical events. Specifically, the paper investigates the behavior of different controllers when they have to keep the average response time of a cloud application below a certain threshold by acting on the probability of serving requests with optional computations disabled, where the pressure exerted by each request on the infrastructure is diminished, at the expense of user experience., Issue: 3

Model-based control improves robot performance provided that the dynamics parameters are estimated accurately. However, some of the model parameters change with time, e.g. friction parameters and unknown payload. Particularly, off-line identification approaches omit the payload estimation (due to practical reasons). Adaptive control copes with some of these structural uncertainties. Thus, this work implements an adaptive control scheme for a 3-DOF parallel manipulator. The controller relies on a novel relevant-parameter dynamic model that permits to study the cases in where the uncertainties affect: (1) rigid body parameters, (2) friction parameters, (3) actuator dynamics, and (4) a combination of the former cases. The simulations and experiments verify the performance of the proposed controller. The control scheme is implemented on the modular programming environment Open Robot Control Software (OROCOS). Finally, an experimental setup evaluates the controller performance when the robot handles a payload.

Lightweight, compliant actuators are particularly desirable in robotic systems intended for interaction with humans. Pneumatic artificial muscles (PAMs) exhibit these characteristics and are capable of higher specific work than comparably-sized hydraulic actuators and electric motors. The objective of this work is to develop a control algorithm that can smoothly and accurately track the desired motions of a manipulator actuated by pneumatic artificial muscles. The manipulator is intended for lifting humans in nursing assistance or casualty extraction scenarios; hence, the control strategy must be capable of responding to large variations in payload over a large range of motion. The present work first investigates the feasibility of two output feedback controllers (proportional-integral-derivative and fuzzy logic), but due to the limitations of pure output feedback control, a model-based feedforward controller is developed and combined with output feedback to achieve improved closed-loop performance. The model upon which the controller is based incorporates the internal airflow dynamics, the physical parameters of the pneumatic muscles and the manipulator dynamics. Simulations were performed in order to validate the control algorithms, guide controller design and predict optimal gains. Using real-time interface software and hardware, the controllers were implemented and experimentally tested on the manipulator, demonstrating the improved capability.

This paper establishes the method of controlled Lagrangian particle tracking (CLPT) to analyze the offsets between physical positions of marine robots in the ocean and simulated positions of controlled particles in an ocean model. The offset, which we term the CLPT error, demonstrates distinguished characteristics not previously seen in drifters and floats that cannot be actively controlled. The CLPT error growth over time is exponential until it reaches a turning point that only depends on the resolution of the ocean model. After this turning point, the error growth slows down significantly to polynomial functions of time. In the ideal case, a theoretical upper threshold on exponential growth of CLPT error can be derived. These characteristics are proved theoretically, verified via simulation, and justified with ocean experimental data. The method of CLPT may be applied to improve the accuracy of ocean circulation models and the performance of navigation algorithms for marine robots. © 2013 IEEE.