D. Tech. (Department of Information Technology, Faculty of Applied and Computer Sciences), Vaal University of Technology. / Simultaneous Localization and Mapping (SLAM) is a significant problem that has been extensively researched in robotics. Its contribution to autonomous robot navigation has attracted researchers towards focusing on this area. In the past, various techniques have been proposed to address SLAM problem with remarkable achievements but there are several factors having the capability to degrade the effectiveness of SLAM technique. These factors include environmental noises (light intensity and shadow), dynamic environment, kidnap robot and computational cost. These problems create inconsistency that can lead to erroneous results in implementation. In the attempt of addressing these problems, a novel SLAM technique Known as DIK-SLAM was proposed. The DIK-SLAM is a SLAM technique upgraded with filtering algorithms and several re-modifications of Monte-Carlo algorithm to increase its robustness while taking into consideration the computational complexity. The morphological technique and Normalized Differences Index (NDI) are filters introduced to the novel technique to overcome shadow. The dark channel model and specular-to-diffuse are filters introduced to overcome light intensity. These filters are operating in parallel since the computational cost is a concern. The re-modified Monte-Carlo algorithm based on initial localization and grid map technique was introduced to overcome the issue of kidnap problem and dynamic environment respectively. In this study, publicly available dataset (TUM-RGBD) and a privately generated dataset from of a university in South Africa were employed for evaluation of the filtering algorithms. Experiments were carried out using Matlab simulation and were evaluated using quantitative and qualitative methods. Experimental results obtained showed an improved performance of DIK-SLAM when compared with the original Monte Carlo algorithm and another available SLAM technique in literature. The DIK-SLAM algorithm discussed in this study has the potential of improving autonomous robot navigation, path planning, and exploration while it reduces robot accident rates and human injuries.

Portfolio optimisation has a number of constraints resulting from some practical matters and regulations. The closed-form mathematical solution of portfolio optimisation problems usually cannot include these constraints. Exhaustive search to reach the exact solution can take prohibitive amount of computational time. Portfolio optimisation models are also usually impaired by the estimation error problem caused by lack of ability to predict the future accurately. A number of Multi-Objective Genetic Algorithms are proposed to solve the problem with two objectives subject to cardinality constraints, floor constraints and round-lot constraints. Fuzzy logic is incorporated into the Vector Evaluated Genetic Algorithm (VEGA) to but solutions tend to cluster around a few points. Strength Pareto Evolutionary Algorithm 2 (SPEA2) gives solutions which are evenly distributed portfolio along the effective front while MOGA is more time efficient. An Evolutionary Artificial Neural Network (EANN) is proposed. It automatically evolves the ANN¿s initial values and structures hidden nodes and layers. The EANN gives a better performance in stock return forecasts in comparison with those of Ordinary Least Square Estimation and of Back Propagation and Elman Recurrent ANNs. Adaptation algorithms for selecting a pair of forecasting models, which are based on fuzzy logic-like rules, are proposed to select best models given an economic scenario. Their predictive performances are better than those of the comparing forecasting models. MOGA and SPEA2 are modified to include a third objective to handle model risk and are evaluated and tested for their performances. The result shows that they perform better than those without the third objective.

<p> This study focuses on developing a new online recursive numerical algorithm for a coupled nonlinear inverse heat conduction-microstructure problem. This algorithm is essential in identifying, designing and controlling many industrial applications such as the quenching process for heat treating of materials, chemical vapor deposition and industrial baking. In order to develop the above algorithm, a systematic four stage research plan has been conducted. </P> <p> The first and second stages were devoted to thoroughly reviewing the existing inverse heat conduction techniques. Unlike most inverse heat conduction solution methods that are batch form techniques, the online input estimation algorithm can be used for controlling the process in real time. Therefore, in the first stage, the effect of different parameters of the online input estimation algorithm on the estimate bias has been investigated. These parameters are the stabilizing parameter, the measurement errors standard deviation, the temporal step size, the spatial step size, the location of the thermocouple as well as the initial assumption of the state error covariance and error covariance of the input estimate. Furthermore, three different discretization schemes; namely: explicit, implicit and Crank-Nicholson have been employed in the input estimation algorithm to evaluate their effect on the algorithm performance. </p> <p> The effect of changing the stabilizing parameter has been investigated using three different forms of boundary conditions covering most practical boundary heat flux conditions. These cases are: square, triangular and mixed function heat fluxes. The most important finding of this investigation is that a robust range of the stabilizing parameter has been found which achieves the desired trade-off between the filter tracking ability and its sensitivity to measurement errors. For the three considered cases, it has been found that there is a common optimal value of the stabilizing parameter at which the estimate bias is minimal. This finding is important for practical applications since this parameter is usually unknown. Therefore, this study provides a needed guidance for assuming this parameter. </p> <p> In stage three of this study, a new, more efficient direct numerical algorithm has been developed to predict the thermal and microstructure fields during quenching of steel rods. The present algorithm solves the full nonlinear heat conduction equation using a central finite-difference scheme coupled with a fourth-order Runge-Kutta nonlinear solver. Numerical results obtained using the present algorithm have been validated using experimental data and numerical results available in the literature. In addition to its accurate predictions, the present algorithm does not require iterations; hence, it is computationally more efficient than previous numerical algorithms. </p> <p> The work performed in stage four of this research focused on developing and applying an inverse algorithm to estimate the surface temperatures and surface heat flux of a steel cylinder during the quenching process. The conventional online input estimation algorithm has been modified and used for the first time to handle this coupled nonlinear problem. The nonlinearity of the problem has been treated explicitly which resulted in a non-iterative algorithm suitable for real-time control of the quenching process. The obtained results have been validated using experimental data and numerical results obtained by solving the direct problem using the direct solver developed in stage three of this work. These results showed that the algorithm is efficiently reconstructing the shape of the convective surface heat flux. </p> / Thesis / Doctor of Philosophy (PhD)

This project is about trying to apply machine learning theories on a selection of data points in order to see if an improvement of current methodology within stress detection and measure selecting could be applicable for the company Linkura AB. Linkura AB is a medical technology company based in Linköping and handles among other things stress measuring for different companies employees, as well as health coaching for selecting measures. In this report we experiment with different methods and algorithms under the collective name of Unsupervised Learning, to identify visible patterns and behaviour of data points and further on we analyze it with the quantity of data received. The methods that have been practiced on during the project are “K-means algorithm” and a dynamic hierarchical clustering algorithm. The correlation between the different data points parameters is analyzed to optimize the resource consumption, also experiments with different number of parameters are tested and discussed with an expert in stress coaching. The results stated that both algorithms can create clusters for the risk groups, however, the dynamic clustering method clearly demonstrate the optimal number of clusters that should be used. Having consulted with mentors and health coaches regarding the analysis of the produced clusters, a conclusion that the dynamic hierarchical cluster algorithm gives more accurate clusters to represent risk groups were done. The conclusion of this project is that the machine learning algorithms that have been used, can categorize data points with stress behavioral correlations, which is usable in measure testimonials. Further research should be done with a greater set of data for a more optimal result, where this project can form the basis for the implementations. / Detta projekt handlar om att försöka applicera maskininlärningsmodeller på ett urval av datapunkter för att ta reda på huruvida en förbättring av nuvarande praxis inom stressdetektering och  åtgärdshantering kan vara applicerbart för företaget Linkura AB. Linkura AB är ett medicintekniskt företag baserat i Linköping och hanterar bland annat stressmätning hos andra företags anställda, samt hälso-coachning för att ta fram åtgärdspunkter för förbättring. I denna rapport experimenterar vi med olika metoder under samlingsnamnet oövervakad maskininlärning för att identifiera synbara mönster och beteenden inom datapunkter, och vidare analyseras detta i förhållande till den mängden data vi fått tillgodosett. De modeller som har använts under projektets gång har varit “K-Means algoritm” samt en dynamisk hierarkisk klustermodell. Korrelationen mellan olika datapunktsparametrar analyseras för att optimera resurshantering, samt experimentering med olika antal parametrar inkluderade i datan testas och diskuteras med expertis inom hälso-coachning. Resultaten påvisade att båda algoritmerna kan generera kluster för riskgrupper, men där den dynamiska modellen tydligt påvisar antalet kluster som ska användas för optimalt resultat. Efter konsultering med mentorer samt expertis inom hälso-coachning så drogs en slutsats om att den dynamiska modellen levererar tydligare riskkluster för att representera riskgrupper för stress. Slutsatsen för projektet blev att maskininlärningsmodeller kan kategorisera datapunkter med stressrelaterade korrelationer, vilket är användbart för åtgärdsbestämmelser. Framtida arbeten bör göras med ett större mängd data för mer optimerade resultat, där detta projekt kan ses som en grund för dessa implementeringar.

No description available.

No description available.

Based on the fact that accessibility in rural areas relies heavily on car traffic, call-driven traffic has been used in different regions to improve public transport usage in rural areas, while it has been difficult to maintain due to high maintenance and long waiting time. Over the past decade, a new demand response transit (DRT) service came into light, which combines with new technologies to provide a more attractive and efficient transport service. Now the Public Transport Authorities have the vision to change this situation of call-driven traffic. In this project, Södertälje and Nykvarn rural area was chosen to be the pilot area of the new DRT service, where a network algorithm was designed to support the routing choices of the new mobility service. The objective of the network algorithm is to achieve an optimal route based on the cost function i.e. operational cost and passengers’ ride time. In addition, the network algorithm is able to test different scenarios, in which user-friendly and operator-friendly scenarios were tested. The result has shown that user-friendly scenarios provide a lower passenger ride time and fleet travel time with the same amount of requests. On top of that, several recommendations regarding improving the service design were proposed in order to optimize customer satisfaction and operation cost. / Baserat på faktumet att tillgängligheten på landsbygden är mer eller mindre byggd på biltrafik så har anropstyrd trafik använts i flera regioner för att förbättra kollektivtrafikanvändningen på landsbygden. Systemet är dock gammalt och lett till svårigheter för drift och underhåll samt långa väntetider. Under det senaste decennium har det utvecklats nya mobilitetslösningar som styrs när behovet uppstår (DRT-service), som kombineras med ny teknik för att ge en mer attraktiv och effektiv transport service. Nu har offentliga transportmyndigheter (PTA) visionen att förbättra denna anropstyrda trafik. I detta projektet valdes landsbygden i Södertälje och Nykvarn som pilotområde för den nya DRT-servicen, där nätverksalgoritmen utformades för att kunna stödja de olika val av vägar för nya mobilitetstjänsten. Målet med nätverksalgoritmen är att uppnå en optimal väg baserad på de kostnadsfunktioner t.ex driktkostnader och passagerarens körtid. Dessutom testas nätverksalgoritmen i olika scenarier, vilket är användarvänliga och driftvänlga scenarier. Resultatet visade att användarvänlga scenarier ger en mindre passagerartid och fordonets resetid gentemot samma mängd av förfrågningar. Dessutom gavs rekomendationer angående hur man kan förbättra tjänstedesignen för att optmiera kundnöjdhet och driftkostnad.

Decline in both visual acuity and visual performance is a fact of life for older people and their increasing share of the population requires that buildings be designed with their visual needs in mind. As their field of vision decreases, people find it harder to identify an objects location, distance, and orientation. Elderly people with vision impairments usually find it harder to perform daily activities such as navigation through indoor spaces. Functional vision can be improved by modifying the design of spaces, for example, with better lighting. However, architects typically do not know how to take the needs of the visually impaired into account in their design process, or simply do not think of doing so. The researcher designed and feasibility-tested a prototype person-centered tool to help architects judge how appropriate a designed space will be for visually impaired people. The study was conducted as a qualitative mixed-methodology research analysis. The researcher used knowledge from literature interpretation to rationalize the development of a person-centered prototype. The researcher immersed design PhD students and vision science experts to inform the prototyping process. Along with an expert group of design and vision science professionals, the researcher beta-tested the prototype during a mock design-process scenario. The researcher also selected a small group of industry experts to participate in open-ended interviews on post-use demonstrations to qualitatively triangulate the findings on the prototypes usability. The study summarizes the feasibility including the challenges of using the prototype for professional purposes and suggests improvement. / Ph. D.

<p dir="ltr">In multiplayer games, a collection of self-interested players aims to optimize their individual cost functions in a non-cooperative manner. The cost function of each player depends not only on its own actions but also on the actions of others. In addition, players' actions may also collectively satisfy some global constraints. The study of this problem has grown immensely in the past decades with applications arising in a wide range of societal systems, including strategic behaviors in power markets, traffic assignment of strategic risk-averse users, engagement of multiple humanitarian organizations in disaster relief, etc. Furthermore, with machine learning models playing an increasingly important role in practical applications, the robustness of these models becomes another prominent concern. Investigation into the solutions of multiplayer games and Nash equilibrium problems (NEPs) can advance the algorithm design for fitting these models in the presence of adversarial noises. </p><p dir="ltr">Most of the existing methods for solving multiplayer games assume the presence of a central coordinator, which, unfortunately, is not practical in many scenarios. Moreover, in addition to couplings in the objectives and the global constraints, all too often, the objective functions contain uncertainty in the form of stochastic noises and unknown model parameters. The problem is further complicated by the following considerations: the individual objectives of players may be unavailable or too complex to model; players may exhibit reluctance to disclose their actions; players may experience random delays when receiving feedback regarding their actions. To contend with these issues and uncertainties, in the first half of the thesis, we develop several algorithms based on the theory of operator splitting and stochastic approximation, where the game participants only share their local information and decisions with their trusted neighbors on the network. In the second half of the thesis, we explore the bandit online learning framework as a solution to the challenges, where decisions made by players are updated based solely on the realized objective function values. Our future work will delve into data-driven approaches for learning in multiplayer games and we will explore functional representations of players' decisions, in a departure from the vector form. </p>

[pt] Atualmente as Unidades de Saúde, em um grande número de países do mundo, apresentam demandas de serviços que superam suas capacidades reais. Por esta razão, o surgimento das listas de espera é inevitável. Preparar o planejamento das mesmas, de modo otimizado resulta, portanto, em um grande desafio, devido à quantidade de recursos que devem ser considerados. O caso particular dos procedimentos cirúrgicos é particularmente crítico pela quantidade de recursos que se precisam para a realização do mesmo. Poucos projetos têm sido desenvolvidos para a gestão completa dessas listas. O trabalho desenvolvido nesta Dissertação propõe o uso de um modelo, baseado em algoritmos genéticos com inspiração quântica, para a automatização e otimização do planejamento de procedimentos cirúrgicos eletivos. Este modelo, denominado Algoritmo Evolucionário com Inspiração Quântica para a Área de Saúde (AEIQ-AS), além de alocar os pacientes e os recursos necessários para que o processo cirúrgico seja exitoso, procura reduzir o tempo total para que todas as cirurgias sejam realizadas. Este trabalho apresenta também uma ferramenta que permite a modelagem, de modo simplificado, de uma Unidade Cirúrgica de Saúde. Esta ferramenta possibilita a realização de simulações com o objetivo de ver o efeito de diferentes configurações dos recursos nas Unidades de Saúde. Para a validação do modelo proposto foi criada, de modo artificial e fazendo uso da ferramenta de simulação, uma lista de espera de 2000 cirurgias. Caso as cirurgias fossem realizadas seguindo a ordem de chegada, seriam necessárias pouco mais de 37 semanas e teria 1066 operações fora do prazo. Foram feitos vários experimentos onde se buscava a otimização destes valores. Esta busca foi feita, primeiramente, tomando em consideração só um dos parâmetros e a continuação eles em conjunto. Na primeira abordagem o AEIQ-AS consegue a realização das mesmas cirurgias em aproximadamente 31 semanas. Assim, observa se que há uma redução de aproximadamente 16,25 porcento do tempo. O número de operações fora do prazo, por sua vez, foi reduzido pelo modelo para 927 (13,04 porcento). Na abordagem simultânea, o AEIQ-AS, consegue uma diminuição do tempo total de alocação em 16,22 porcento e o número de operações fora do prazo em 9,76 porcento. Foram feitas, também, várias simulações da Unidade de Saúde mantendo as caraterísticas da lista de cirurgias para ver seu efeito no tempo total de alocação de todos os processos cirúrgicos. / [en] Currently, Health Units in a large number of countries in the world present service demand that exceed their real capacities. For this reason, is inevitable the emergence of the waiting lists. To prepare the planning of this in an optimized manner results in a substantial challenge due to the number of resources that should be considered. The case of chirurgical procedures is particularly critical by the number of resources needed for their realization. A small quantity of projects has been developed to fully manage these lists. The work developed in this Dissertation proposes the use of a model based on evolutionary algorithms with quantum inspiration for the automation and optimization of the planning of elective chirurgical procedures. This model, denominated Evolutionary Algorithm with Quantum Inspiration for the Health Field (AEIQ-AS), beyond patients and necessary resources for the successful completion of the chirurgical procedure allocation, pursue the reduction of the total time of realization of all the surgeries. The work presents also a tool that allows the modeling, in a simplified manner, of a Chirurgical Health Unit. This tool enables the realization of simulations with the objective of seeing the effect of different configurations of the resources in the Health Units. To validate the proposed model was created, in artificial mode and employing the simulation tool, a waiting list of 2000 surgeries. In case that the surgeries were realized following the arrival order, will be needed a little more than 37 weeks and will have 1066 surgeries out of time. Several experiments were conducted in order to optimize these values. This search was executed, firstly, considering only one of the parameters and, in continuation, all together. In the first approach, the AEIQ-AS obtains the realization of the same surgeries in approximately 16,25 percent of the time. The number of operations out of time was reduced by the model to 927 (13,04 percent). In the simultaneous approach, the AEIQAS achieves a decrease of the allocation total time in 16,22 percent and the number of operations out of time in 9,76 percent. It were done, also, several simulations of the Health Unit maintaining the characteristics of the surgeries list in order to look the effect in the allocation total time of all the chirurgical procedures.