Induction motor is the most popular machine in the industry. It is used extensively in mechanical plants, and it is un- avoidable to have the motor¡¦s electrical and mechanical faults due to continuously operating throughout the year. Faults of motors do not only cause the production line to shut down but also imperil the personnel security. A suitable motor maintenance schedule will be a needed to decrease the machine down time. However, major investment might take up to 90% for equipment, and it would be helpful to have a practicable low-cost supervisory scheme on maintenance. If the faults of machine can be detected correctly and effectively, the maintenance efficiency and dependability could be increased greatly. In the past, researches on fault recognition for Induction motors only concentrated on Spectrum analysis with amplitudes based on a constant load. However, the frequency and amplitude of the spectrum analyzed under different fault conditions are also affected significantly by load variations. Using spectrum amplitudes to recognize motor faults is not sufficient in a practical system. Various types of faults and load conditions will influence the spectrum structure. In order to recognize faults under various load conditions, we must consider band shift and amplitude variation as two major factors. In this paper, we use the methods of frequency axis adjustment, load interval and feature exaction to solve the band shift and amplitude variation problems respectively. After the above-mentioned procedures, efficient features are obtained. We use the Back Propagation Neural Network (BPNN) and General Regression Neural Network (GRNN) to train and recognize fault conditions.

Conceptual estimation techniques play an important role in determining the approximate costs of construction projects especially during feasibility stages. Moreover, pre-design estimates are also crucial for the contractors. With the help of the conceptual predictions companies can determine approximate project costs and can gain several advantages before tendering phase. The main objective of this thesis is to focus on modeling of quantities instead of costs and to develop quantity take-off models for pre-design cost estimation of bridge projects. Majority of the existing studies focus on modeling of costs for conceptual cost estimation. This study includes modeling of the quantity take off items in a specific single span highway bridge using three different techniques namely, linear regression, neural network and case based reasoning. During this study 40 single span highway bridge projects whose owner is Republic of Turkey General Directorate of Railways, Ports and Airports Constructions were investigated and models for each work item were developed. Then by integrating the quantity take off estimations with unit costs, total project costs were calculated. As a result by evaluating the prediction performance of the models, comparison of the methods was achieved. Results are discussed along with the advantages of the proposed method for conceptual cost estimation of bridge projects.

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

The application area of Fifth Generation New Radio (5G-NR) called Ultra-Reliable and Low-Latency Communication (URLLC) requires a reliability, the probability of receiving and decoding a data packet correctly, of 1 - 10^5. For this requirement to be fulfilled in a resource-efficient manner, it is necessary to have a good estimation of extremely low quan- tiles of the channel quality distribution, so that appropriate resources can be distributed to users of the network system.  This study proposes and evaluates two methods for estimating extreme quantiles of the downlink channel quality distribution, linear quantile regression and Quantile Regression Neural Network (QRNN). The models were trained on data from Ericsson’s system-level radio network simulator, and evaluated on goodness of fit and resourcefulness. The focus of this study was to estimate the quantiles 10^2, 10^3 and 10^4 of the distribution.  The results show that QRNN generally performs better than linear quantile regression in terms of pseudoR2, which indicates goodness of fit, when the sample size is larger. How- ever, linear quantile regression was more effective for smaller sample sizes. Both models showed difficulty estimating the most extreme quantiles. The less extreme quantile to esti- mate, the better was the resulting pseudoR2-score. For the largest sample size, the resulting pseudoR2-scores of the QRNN was 0.20, 0.12 and 0.07, and the scores of linear quantile regression was 0.16, 0.10 and 0.07 for the respective quantiles 10^2, 10^3 and 10^4.  It was shown that both evaluated models were significantly more resourceful than us- ing the average of the 50 last measures of channel quality subtracted with a fixed back-off value as a predictor. QRNN had the most optimistic predictions. If using the QRNN, theo- retically, on average 43% more data could be transmitted while fulfilling the same reliability requirement than by using the fixed back-off value.

Autonomous cars involve problems with control and planning. In thispaper, we implement and evaluate an autonomous agent based ona Monte-Carlo Tree Search in continuous action space. To facilitatethe algorithm, we extend an existing simulation framework and usea GPU for faster calculations. We compare three action generatorsand two rewards functions. The results show that MCTS convergesto an effective driving agent in static environments. However, it onlysucceeds at driving slow speeds in real-time. We discuss the problemsthat arise in dynamic and static environments and look to future workin improving the simulation tool and the MCTS algorithm. See code, https://github.com/felrock/PyRacecarSimulator

This degree project addresses poaching challenges by employing predictive analysis of animal movements and their correlation with the dynamic environment using a machine learning approach. The goal is to provide accurate predictions of animal movements, enabling rangers to intercept potential threats and safeguard wildlife from snares. A wide analysis considers previous studies on animal movements and both animal and environment data availability. To efficiently represent the dynamic environment and correlate it with animal movement data, accurate matching of environment variables to each animal measurement is crucial. We selected multiple environment datasets to capture a sufficient amount ofenvironmental properties. Due to practical constraints, daily representation of the environment is not achievable, and weekly mean or monthly mode values are used instead. Data insights are obtained through the training of a regression neural network using the filtered environmental and animal movement data. The results highlight the significant role ofenvironmental features in predicting animal movements, emphasizing their importance for accurate predictions. Despite some offset and few erroneous predictions, a strong similarity between animal predicted trajectory and animal true trajectory was achieved, indicating that the model is capable to capture general patterns and to correctly tune in predictions of detailed movements as well. The overall offset of the trajectories is still a weak point of this model, but it may just indicate the presence of some underlying systematic error that can be corrected through further work. The integration of such a developed prediction model into existing frameworks could assist law enforcingauthorities in preventing poaching activities.

No / This paper presents an investigation into the challenges in implementing a hard real-time optimal non-stationary system using general regression neural network (GRNN). This includes investigation into the dynamics of the problem domain, discretisation of the problem domain to reduce the computational complexity, parameters selection of the optimization algorithm, convergence guarantee for real-time solution and off-line optimization for real-time solution. In order to demonstrate these challenges, this investigation considers a real-time optimal missile guidance algorithm using GRNN to achieve an accurate interception of the maneuvering targets in three-dimension. Evolutionary Genetic Algorithms (GAs) are used to generate optimal guidance training data set for a large missile defense space to train the GRNN. The Navigation, Constant of the Proportional Navigation Guidance and the target position at launching are considered for optimization using GAs. This is achieved by minimizing the. miss distance and missile flight time. Finally, the merits of the proposed schemes for real-time accurate interception are presented and discussed through a set of experiments. (C) 2012 Elsevier Ltd. All rights reserved.

This thesis discusses the modeling of a micro-grid with photovoltaic (PV)-wind-fuel cell (FC) hybrid energy system and its operations. The system consists of the PV power, wind power, FC power, static var compensator (SVC) and an intelligent power controller. Wind and PV are primary power sources of the system, and an FC-electrolyzer combination is used as a backup and a long-term storage system. A simulation model for the micro-grid control of hybrid energy system has been developed using MATLAB/Simulink. A SVC was used to supply reactive power and regulate the voltage of the hybrid system. To achieve a fast and stable response for the real power control, the intelligent controller consists of a Radial Basis Function Network-Sliding Mode Control (RBFNSM) and a General Regression Neural Network (GRNN) for maximum power point tracking (MPPT). The pitch angle of wind turbine is controlled by RBFNSM, and the PV system uses GRNN, where the output signal is used to control the DC/DC boost converters to achieve the MPPT.

The frequency estimation is a problem widely studied in many fields including electric power systems. Several methods have been proposed for this purpose, and most of them perform well when the signal is not distorted by harmonics or noises. This paper presents two new methods based on Artificial Neural Networks for frequency estimation. Both use Clarck s transform to generate a phasor that represent the system s signal. In the first methodology this phasor is normalized and feeds the Generalized Regression Neural Network, that ponders the values. At the end it s obtained a phasor where noisy and harmonics are attenuated. The neural network output is then used to calculate the electrical system frequency. Otherwise, the second methodology uses the Adaptive Linear Neural Network. This work tested also various methodologies of frequency estimation proposed in other knowledge fields such as radar, sonar, communications, biomedicine and aviation however with electrical power systems signals. These methods are: Lavopa (proposed by Lavopa et al. 2007), Quinn (proposed by Quinn, 1994), Jacobsen (proposed by Jacobsen e Kootsookos, 2007), Candan (proposed by Candan, 2011), Macleod (proposed by Macleod, 1998), Aboutanios (proposed by Aboutanios, 2004), Mulgrew (proposed by Aboutanios e Mulgrew, 2005), Ferreira (proposed by Ferreira 2001) e DPLL (proposed by Sithamparanathan, 2008). With the exception of DPLL the remaining methods are based on the Discrete Fourier Transform and seek the spectrum frequency peak to than find the fundamental frequency. The nine methodologies are compared with the proposed methods and with the commonly techniques used or studied for electric power systems. Tests include noisy signals, harmonics, sub-harmonics, frequency variations on step, ramp and sinusoidal, also variations on voltage and phase are considered. The tests also include a simulated signal where a load block is inserted and immediately after removed from the system. At the end a comparison is made between the techniques, been able to point each technique advantage and disadvantage trough the comparison identify the best methods to be applied on electrical power systems. / A estimação de frequência é um problema muito estudado em diversas áreas, dentre elas a dos sistemas elétricos de potência. Inúmeras metodologias foram propostas para esse fim, sendo que a maioria delas apresenta bom desempenho quando o sinal não está distorcido por componentes harmônicas ou ruídos. Este trabalho propõe duas novas metodologias fundamentadas em Redes Neurais Artificiais, de modo a estimar a frequência. Elas utilizam a transformada de Clarck para gerar um fasor que representa o sinal trifásico do sistema. Na primeira metodologia, esse fasor é normalizado e alimenta a Rede Neural de Regressão Generalizada, que faz a ponderação dos valores. Ao final, obtém-se um fasor em que ruídos e harmônicas são atenuados. A saída da rede neural é, então, utilizada para o cálculo da frequência do sistema elétrico. A segunda metodologia utiliza a Rede Neural Linear Adaptativa. Neste trabalho, também são testadas, para uso em sistemas elétricos de potência, diversas metodologias propostas em outras áreas de conhecimento, tais como radar, sonar, comunicação, biomedicina e aviação. São elas: Lavopa (proposta por Lavopa et al. 2007), Quinn (proposta por Quinn, 1994), Jacobsen (proposta por Jacobsen e Kootsookos, 2007), Candan (proposta por Candan, 2011), Macleod (proposta por Macleod, 1998), Aboutanios (proposta por Aboutanios, 2004), Mulgrew (proposta por Aboutanios e Mulgrew, 2005), Ferreira (proposta por Ferreira 2001) e DPLL (proposta por Sithamparanathan, 2008). Com exceção da DPLL, os demais métodos são fundamentados na transformada discreta de Fourier e buscam encontrar o pico do espectro de frequências, para, então, encontrar a frequência fundamental. As nove metodologias são comparadas juntamente com os métodos propostos e as técnicas já comumente usadas ou estudadas para sistemas elétricos. Os testes incluem sinais com ruídos, harmônicas, sub-harmônicas, variações de frequência em degrau, rampa e senoidal, variações de fase e tensão em degrau. Os testes ainda incluem um sinal provindo de simulação em que um bloco de carga é inserido e logo após retirado do sistema. Ao final é realizada uma comparação entre as técnicas, sendo possível identificar as vantagens e desvantagens de cada uma e, assim, indicar as melhores a serem usadas em sistemas elétricos de potência.

Tato práce pojednává o technikách behaviorálního modelování pro speciální úlohy v elektromagnetismu, které je možno formulovat jako problém aproximace, klasifikace, odhadu hustoty pravděpodobnosti nebo kombinatorické optimalizace. Zkoumané methody se dotýkají dvou základních problémů ze strojového učení a combinatorické optimalizace: ”bias vs. variance dilema” a NP výpočetní komplexity. Boltzmanův stroj je v práci navržen ke zjednodušování komplexních impedančních sítí. Bayesovský přístup ke strojovému učení je upraven pro regularizaci Parzenova okna se snahou o vytvoření obecného kritéria pro regularizaci pravděpodobnostní a regresní neuronové sítě.