Data mining plays an important role in a variety of fields including bioinformatics, multimedia, business intelligence, marketing, and medical diagnosis. Analysis of today’s huge and complex data involves several data mining algorithms including clustering and classification. The computational complexity of machine learning and data mining algorithms, that are frequently used in today’s applications such as embedded systems, makes the design of efficient hardware architectures for these algorithms a challenging issue for the development of such systems. The aim of this work is to optimize the performance of hardware acceleration for data mining applications in terms of speed and area. Most of the previous accelerator architectures proposed in the literature have been obtained using ad hoc techniques that do not allow for design space exploration, some did not consider the size (number of samples) and dimensionality (number of features in each sample) of the datasets. To obtain practical architectures that are amenable for hardware implementation, size and dimensionality of input datasets are taken into consideration in this work. For one-dimensional data, algorithm-level optimizations are investigated to design a fast and area-efficient hardware accelerator for clustering one-dimensional datasets using the well-known K-Means clustering algorithm. Experimental results show that the optimizations adopted in the proposed architecture result in faster convergence of the algorithm using less hardware resources while maintaining the quality of clustering results. The computation of similarity distance matrices is one of the computational kernels that are generally required by several machine learning and data mining algorithms to measure the degree of similarity between data samples. For these algorithms, distance calculation is considered a computationally intensive task that accounts for a significant portion of the processing time. A systematic methodology is presented to explore the design space of 2-D and 1-D processor array architectures for similarity distance computation involved in processing datasets of different sizes and dimensions. Six 2-D and six 1-D processor array architectures are developed systematically using linear scheduling and projection operations. The obtained architectures are classified based on the size and dimensionality of input datasets, analyzed in terms of speed and area, and compared with previous architectures in the literature. Motivated by the necessity to accommodate large-scale and high-dimensional data, nonlinear scheduling and projection operations are finally introduced to design a scalable processor array architecture for the computation of similarity distance matrices. Implementation results of the proposed architecture show improved compromise between area and speed. Moreover, it scales better for large and high-dimensional datasets since the architecture is fully parameterized and only has to deal with one data dimension in each time step. / Graduate / 2019-12-31

O computador vem ocupando atualmente um papel de crescente destaque na vida do homem moderno. Nesse contexto, também a área de projeto e, mais particularmente, a do projeto de edifícios, encontram nele uma ferramenta de inestimável valor. Seu atual nível de utilização nesse setor, no entanto, não tem se aproveitado de todo o potencial oferecido. Um dos motivos que justificam tal fato está no desconhecimento, por parte dos projetistas, das reais possibilidades de emprego dos computadores na área do projeto de edifícios. Nesse sentido, o objetivo dessa Dissertação é contribuir para a organização dos conceitos necessários a tais empregos, bem como montar um painel sobre as perspectivas de utilização dessa poderosa ferramenta na área. / The computer is playing an increasing role in the life of the modern man. With this point of view the design and more specifically the building design can make the computer a valuable tool. However, despite the potential of this equipment, it has not been yet properly used. This happens because the designers do not know the actual possibilities of the use of computers for the building designs. This Dissertation aims to organize the concepts related with this possibilities and also to bring out the prospective uses of this power full tool in this area.

Esta pesquisa investiga as possibilidades metodológicas do campo do design que possam lidar eficientemente com sistemas complexos que servem de contexto a vários problemas contemporâneos. Entendemos que os problemas propostos à área projetual estão se tornando progressivamente mais complexos, sem que tenha havido uma contrapartida na sofisticação da reflexão dos métodos projetuais. Partimos então de uma pesquisa das correntes metodológicas de projeto -- especificamente da corrente Design Methods anglo-americana, que desde sua criação nos anos 1960 colocou a necessidade de novos métodos para o design contemporâneo -- que pudessem dar conta de novos níveis de complexidade. Num segundo momento, procuramos fundamentação na teoria dos sistemas, a respeito dos fenômenos típicos dos sistemas complexos -- emergência, robustez, grandes eventos --, suas características e desdobramentos para o mapeamento das soluções dos problemas de design. A partir desta fundamentação identificamos estratégias para o tratamento de problemas complexos. Também procuramos analisar e inferir, das várias iniciativas da web colaborativa, os elementos que geram soluções para problemas de grande complexidade. No decorrer da pesquisa foi possível perceber que as iniciativas existentes na web que tratam de problemas complexos se utilizam de certas estratégias que possibilitam abordagens de projeto mais efetivas. Ao final do estudo, procuramos elaborar uma síntese dessas estratégias e lançar sugestões de abordagem que possam ajudar projetistas no enfrentamento dos problemas complexos típicos de nossos tempos. / This research investigates the possibilities of a design methodology that can deal with complex systems that serve as the backdrop to many contemporary problems. We think that the problems posed to design field are becoming increasingly complex, without any correspondent counterpart in the sophistication in design methods. Therefore, we started with a survey in design methodological theories -- specifically the Anglo-American Design Methods movement, which since its creation in the 1960s put the need for new methods to contemporary design -- that could deal with new levels of complexity. As a second step, we seek grounding in systems theory, about the typical complex systems phenomena -- emergence, robustness, major events -- its features and developments for the mapping of solutions of design problems. From this foundation we identified different strategies to solve complex problems. We also seek to analyze and infer, based on the various collaborative web initiatives, the elements that generate solutions to problems of great complexity. During the research it was revealed that existing web initiatives that address complex problems make use of some strategies that enable more effective design approaches. At the end of this work, we sought to develop an overview of these strategies and proposed suggested approaches that can help designers in addressing the complex problems typical of our times.

This thesis examines the use of a structured design methodology in the design of asynchronous circuits so that high level constructs can be specified purely in terms of signal exchanges and without the intrusion of lower level concepts. Trace theory is used to specify a multi-processor Forth machine at a high level then part of the design is further elaborated using trace theory operations to (insure that the behaviours of the lower level constructs will combine to give the high level specified behaviour without locking or other hazards. A novel form of threaded language to take advantage of the machine architecture is developed. At suitable points the design is tested by simulation. The stack element which is designed is reduced to an electric circuit which is itself tested by simulation to verify the design.

The simulation of surgical procedures, in the case of hard tissues such as bone or teeth milling, using a haptic milling surgery simulator requires a haptic device which can provide high stiffness and transparency. To mimic a real milling process of hard tissue, such as for example creating a narrow channel or cavity, the simulator needs to provide force/torque feedback in 5–6 degrees of freedom (DOF). As described in this thesis, research has been performed to develop and optimize a haptic device that can provide high stiffness and force/torque capabilities to facilitate haptic interaction with stiff tissues.  The main contributions of this thesis are: (i) The use of a model-based design methodology for the design of haptic devices.  The proposed methodology is applied to a case study, i.e. the design and optimization of a haptic device based on parallel kinematics. Device requirements were elicited through dialogues with a prospective user from a neurosurgery clinic. In the conceptual design phase, different parallel concepts have been investigated and analyzed based on functional qualities such number of degrees of freedom, workspace size and force/torque capabilities. This analysis led to the selection of a specific 6 DOF kinematic structure for which dimension synthesis was performed including multi-objective optimization followed by control synthesis. Finally, a device prototype was realized and its performance verified. (ii) Optimization of the device for best kinematic and dynamic performance. For optimization, performance indices such as workspace-to-footprint ratio, kinematic isotropy and inertial indices were used. To cope with the problem of non-uniform units in the components of the Jacobian matrix, various normalization techniques were investigated. A new multi-objective optimization function is introduced to define the optimization problem, which is then resolved using multi-objective genetic algorithms. A sensitivity analysis of the performance indices against each design parameter is performed, as a basis for selecting a final set of design parameter values. (iii) A control strategy is investigated to achieve high transparency and stability of the device. The control strategy is based on careful analysis of the dynamics of the haptic device, computed torque feed-forward control and force control based on current feedback. (iv) Finally, experiments both separately in the lab and by using the device in a haptic milling surgery simulator were performed. Results from a face validity study performed in collaboration with orthopedists verify that the new haptic device enables high-performance force and torque feedback for stiff interactions. / QC 20120302

Traditional methods for mechatronics design are often based on a sequential approach, where the mechanical structure is designed first, and then fitted with off-the-shelf electric motors, drive electronics, gearheads and sensors. Finally a control system is designed and optimized for the already existing physical system. Such a design method, that doesn’t consider aspects from a control point of view during the design of the physical system, is unlikely to result in a system with optimal control performance. Furthermore, to separately design and optimize each of the physical components will, from a global perspective, generally not result in a system that is optimal from a weight, size or cost perspective. In order to reach the optimal design of an integrated mechatronic system (mechatronic module) it is necessary to treat the system as a whole, considering aspects from all involved engineering domains concurrently. In this thesis such an approach to integrated design of mechatronic servo systems is presented. A design methodology that considers the simultaneous design of the electric machine, gearhead, machine driver and control system, and therefore enables global optimization, has been developed. The target of the design methodology is conceptual design and evaluation. It is assumed that the load to be driven by the servo system is known and well defined, a load profile describing the wanted load motion and the corresponding torque, is required as input. The methodology can then be used to derive the lightest or smallest possible system that can drive the specified load. Furthermore, the control performance is evaluated and optimized, such that the physical system design and the controller design are integrated. The methodology is based on modelling and simulation. Two types of component models have been developed, static and dynamic models. The static models describe relations between the parameters of the physical components, for example a component’s torque rating as function of its size. The static models are based on traditional design rules and are used to optimize the physical parts of the system. The dynamic models describe the behaviour of the components and are used for control system design and performance optimization. The gear ratio is identified to be the most central design variable when designing and optimizing electromechanical servo systems. The gear ratio directly affects the required size of the gearhead, electric machine and the machine driver. But it has also large influences on the system’s control performance. It is concluded that high gear ratios generally are better from a control point of view than low ratios. A consequence of this is that it is possible, without compromising the control performance, to use less expensive (less accurate) sensors and microprocessors in high gear ratio servo systems, while low gear ratio systems require more expensive hardware. It is also concluded that it is essential to include all performance limiting phenomena, linear as well as non-linear, in this type of integrated analysis. Using for example a linearized system description for controller design, means that many of the most important couplings between control system and physical system design are overlooked. / QC 20100816

This thesis presents contributions within two domains of manufacturing system development. The first domain is systematic design of product and manufacturing system. The second domain is model based co-operative development between manufacturing companies and manufacturing system suppliers. The axiomatic design framework for design is used to describe the product design and the manufacturing system design. The most important contributions are that product properties and analysis of product function are considered when specifying the manufacturing system. The systematic adaptation of product to the selected manufacturing solution is another contribution. A new concept of manufacturing phase functional requirements has been defined in order to cope with requirements from the manufacturing process on the in process product. The co-operative concept is based on the use of ISO 10303-214 for specification and communication of proposed systems and equipment. The concept has been tested using real project information in prototype implementation / QC 20100913

It was suggested that dirt was a way of falling away from perfection and through introducing or allowing dirt you could avoid the sterile and lifeless. A ephemeral and intangible entity, dirt is, by its nature, an elusive creature. This intangibility encouraged me to draw a clearer definition for myself of what dirt meant. If dirt was not limited to patina and grime how could it be understood in terms of design. This is the essence of my thesis. / Some pictures have been removed for publication.

An epicyclic-type automatic transmission is a device which is connected from the engine to the driven wheels. A Lepelletier automatic transmission consists of a two-DOF epicyclic gear mechanism and a fundamental geared entity, which can provide more speed ratios. For the development of automatic transmission with eight or more speed ratios, an efficient design methodology is urgently needed and still not available. Therefore, the purpose of this article is to develop a systematic method for the design of Lepelletier automatic transmission. First, the features and requirements of the Lepelletier automatic transmissions are investigated and the corresponding design requirements are induced. Next, the theory of the analysis of speed ratios for Lepelletier automatic transmission is derived based on the concept of lever analogy. Then, the design theory of Lepelletier automatic transmission is developed based on the analysis theory. Finally, the complete atlas of six-speed and eight-speed Lepelletier automatic transmissions can be systematically synthesized from the data of three given speed ratios. In this article, the innovative and patentable six-speed and eight-speed Lepelletier automatic transmissions with the simplest structure have been. This proposed design method can also be applied to the development and design of ten-speed Lepelletier automatic transmissions.

In this project the design of a double base for two child restraint seats (CRS) has been developed. The objective of this project is to facilitate the transportation of two children in the rear seats of a car, maximizing the space available in the rear of the car and providing a comfortable movement to the user when placing the CRS baskets into the car. Scientific articles have been analysed providing a reference of the strength limitations and ergonomics aspects of the users as well as misuse of current CRS. Also, ergonomics and safety issues have been studied through these articles in order to fulfil with the current regulations established by the EU standards and the regulation 44 of Economic Commission for Europe. A market research and survey investigations have been done in order to study the features of existing CRS designs, opinion and needs of habitual users. The employed methodology based on a divergent and convergent design process has been carried out through different methods such as Brainstorming, 6-3-5 Method and Dark Horse Prototype in which many ideas, concepts and proposals have been developed and explained. The results of this project are shown and explained through a virtual 3D CAD model with their respective renders, explanations and technical drawings. The use and material selection of the final design are presented as well as the results of an ergonomic study performed by test with a simple physical prototype of the final design. The main outcomes of this project are included in conclusion, discussions and future work sections.