In this research work, an approach for fire detection and estimation robots is presented. The approach is based on type-2 fuzzy logic system that utilizes measured temperature and light intensity to detect fires of various intensities at different distances. Type-2 fuzzy logic system (T2 FLS) is known for not needing exact mathematic model and for its capability to handle more complicated uncertain situations compared with Type-1 fuzzy logic system (T1 FLS). Due to lack of expertise for new facilities, a new approach for training experts’ expertise and setting up T2 FLS parameters from pure data is discussed in this thesis. Performance of both T1 FLS and T2 FLS regarding to same fire detection scenario are investigated and compared in this thesis. Simulation works have been done for fire detection robot of both free space scenario and new facility scenario to illustrate the operation and performance of proposed type-2 fuzzy logic system. Experiments are also performed using LEGO MINDSTROMS NXT robot to test the reliability and feasibility of the algorithm in physical environment with simple and complex situation.

Submitted by Geandra Rodrigues (geandrar@gmail.com) on 2018-01-09T13:48:15Z No. of bitstreams: 1 renanpiazzarolifinottiamaral.pdf: 1172046 bytes, checksum: eb7bf10c813d64fbddcc572d39aecfc5 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2018-01-22T16:10:30Z (GMT) No. of bitstreams: 1 renanpiazzarolifinottiamaral.pdf: 1172046 bytes, checksum: eb7bf10c813d64fbddcc572d39aecfc5 (MD5) / Made available in DSpace on 2018-01-22T16:10:30Z (GMT). No. of bitstreams: 1 renanpiazzarolifinottiamaral.pdf: 1172046 bytes, checksum: eb7bf10c813d64fbddcc572d39aecfc5 (MD5) Previous issue date: 2017-09-01 / - / This thesis presents and discusses improvements in the type-1 and singleton fuzzy logic system for dealing with classification problems. Two training methods are addressed, the scaled conjugate gradient, which uses the second order information approximating the multiplication of the Hessian matrix H by the directional vector v (i.e. Hv), and the same method using the differential operator R {.} to compute the exact value of Hv. Also, in order to adapt the fuzzy model to handle multiclass classification problems, it is developed a novel fuzzy model with a vector as output. All proposals are tested through the performance metrics analysis based on data sets provided by UCI Machine Learning Repository. The reported results show the high convergence speed and better classification rates of the proposed training methods than others presented in the literature. Additionally, the novel fuzzy model has a significant reduction in computational and classifier complexity, especially when the number of classes in classification problem increases.

Increasing demands, shrinking financial and human resources, and increased infrastructure deterioration have made the task of maintaining the infrastructure systems more challenging than ever before. Life-cycle cost analysis (LCCA) is an important tool for transportation infrastructure management, which is used extensively to support project level decisions, and is increasingly being applied to enhance network level analysis. However, traditional LCCA tools cannot practically and effectively utilize expert knowledge and handle ambiguous uncertainties. The main objective of this dissertation was to develop enhanced LCCA models using soft computing (mainly fuzzy logic) techniques. The proposed models use available "real-world" information to forecast life-cycle costs of competing maintenance and rehabilitation strategies and support infrastructure management decisions. A critical review of available soft computing techniques and their applications in infrastructure management suggested that these techniques provide appealing alternatives for supporting many of the infrastructure management functions. In particular, LCCA often utilizes information that is uncertain, ambiguous and incomplete, which is obtained from both existing databases and expert opinion. Consequently, fuzzy logic techniques were selected to enhance life-cycle cost analysis of transportation infrastructure investments because they provide a formal approach for the effective treatment of these types of information. The dissertation first proposes a fuzzy-logic-based decision-support model, whose inference rules can be customized according to agency's management policies and expert opinion. The feasibility and practicality of the proposed model is illustrated by its implementation in a life-cycle cost analysis algorithm for comparing and selecting pavement maintenance, rehabilitation and reconstruction (MR&R) policies. To enhance the traditional probabilistic LCCA model, the fuzzy-logic-based model is then incorporated into the risk analysis process. A fuzzy logic approach for determining the timing of pavement MR&R treatments in a probabilistic LCCA model for selecting pavement MR&R strategies is proposed. The proposed approach uses performance curves and fuzzy-logic triggering models to determine the most effective timing of pavement MR&R activities. The application of the approach in a case study demonstrates that the fuzzy-logic-based risk analysis model for LCCA can effectively produce results that are at least comparable to those of the benchmark methods while effectively considering some of the ambiguous uncertainty inherent to the process. Finally, the research establishes a systematic method to calibrate the fuzzy-logic based rehabilitation decision model using real cases extracted from the Long Term Pavement Performance (LTPP) database. By reinterpreting the model in the form of a neuro-fuzzy system, the calibration algorithm takes advantage of the learning capabilities of artificial neural networks for tuning the fuzzy membership functions and rules. The practicality of the method is demonstrated by successfully tuning the treatment selection model to distinguish between rehabilitation (light overlay) and do-nothing cases. / Ph. D.

Vehicle acoustic signals have long been considered as unwanted traffic noise. In this research acoustic signals generated by each vehicle will be used to detect its presence and classify its type. Circular arrays of microphones were designed and built to detect desired signals and suppress unwanted ones. Circular arrays with multiple rings have an interesting and important property that is constant sidelobe levels. A modified genetic algorithm that can work directly with real numbers is used in the circular array design. It offers more effective ways to solve numerical problems than a standard genetic algorithm. In classifier design two main paradigms are considered: multilayer perceptrons and adaptive fuzzy logic systems. A multilayer perceptron is a network inspired by biological neural systems. Even though it is far from a biological system, it possesses the capability to solve many interesting problems in variety fields. Fuzzy logic systems, on the other hand, were inspired by human capabilities to deal with fuzzy terms. Its structures and operations are based on fuzzy set theory and its operations. Adaptive fuzzy logic systems are fuzzy logic systems equipped with training algorithms so that its rules can be extracted or modified from available numerical data similar to neural networks. Both fuzzy logic systems and multilayer perceptrons have been proved to be universal function approximators. Since there are approximations in almost every stage, both of these system types are good candidates for classification systems. In classification problems unequal learning of each class is normally encountered. This unequal learning may come from different learning difficulties and/or unequal numbers of training data from each class. The classifier tends to classify better for a well-learned class while doing poorly for other classes. Classification costs that may be different from class to class can be used to train and test a classifier. An error backpropagation algorithm can be modified so that the classification costs along with unequal learning factors can be used to control classifier learning during its training phase. / Ph. D.

Dizertační práce se zabývá návrhem integrované metody hodnocení bezpečnosti a spolehlivosti palubních leteckých systém za použití teorie grafů a fuzzy logiky. Navržená integrovaná metoda je univerzálně použitelná v oblasti hodnocení bezpečnosti a spolehlivosti, nicméně je primárně navržená pro použití v oblasti General Aviation a civilních bezpilotních prostředků. Současná podoba hodnocení spolehlivosti je téměř výhradně závislá na úsudku analytika. Použití komerčních softwarových nástrojů pro hodnocení spolehlivosti je extrémně nákladné, přičemž možnost přístupu a úpravy použitých algoritmů je minimální. Současný prudký vývoj palubních letecký systému je spojen s jejich zvyšující se komplexností a sofistikovaností. Integrovaná metoda používá teorii grafů, jako nástroj modelování funkčních závislostí mez jednotily prvky systému. Použití teorie grafu současně umožňuje daný systém analyzovat, hodnotit hustotu vzájemné funkční vazebnosti, identifikovat důsledky případných poruchových stavů. Aplikace fuzzy logiky umožňuje manipulovat s expertní znalostí a stanovit kritičnost daného prvku a systému. Kritičnost prvku zohledňuje pravděpodobnost jeho selhání, možnost detekce dané poruchy, závažnost těchto selhání vzhledem k vlivu na alokované funkce.

In human-robot collaborative (HRC) scenarios where humans and robots work together sharing the same workspace, there is a risk of potential hazard that may occur. In this work, an AI-based risk analysis solution has been developed to identify any condition that may harm a robot and its environment. The information from the risk analysis is used in a risk mitigation module to reduce the possibility of being in a hazardous situation. The goal is to develop safety for HRC scenarios using different AI algorithms and to check the possibilities of improving efficiency of the system without any compromise on the safety. This report presents risk mitigation strategies that were built on top of the robot’s control system and based on the ISO 15066 standard. Each of them used semantic information (scene graph) about the robot’s environment and changed the robot’s movement by scaling speed. The first implementation of risk mitigation strategy used Fuzzy Logic System. This system analyzed the riskiest object’s properties to adjust the speed of the robot accordingly. The second implementation used Reinforcement Learning and considered every object’s properties. Three networks (fully connected network, convolutional neural network, and hybrid network) were implemented to estimate the Qvalue function. Additionally, local and edge computation architecture wereimplemented to measure the computational performance on the real robot. Each model was evaluated by measuring the safety aspect and the performance of the robot in a simulated warehouse scenario. All risk mitigation modules were able to reduce the risk of potential hazard. The fuzzy logic system was able to increase the safety aspect with the least efficiency reduction. The reinforcement learning model had safer operation but showed a more compromised efficiency than the fuzzy logic system. Generally, the fuzzy logic system performed up to 28% faster than reinforcement learning but compromised up to 23% in terms of safety (mean risk speed value). In terms of computational performance, edge computation was performed faster than local computation. The bottleneck of the process was the scene graph generation which analyzed an image to produce information for safety analysis. It took approximately 15 seconds to run the scene graph generation on the robot’s CPU and 0.3 seconds on an edge device. The risk mitigation module can be selected depending on KPIs of the warehouse operation while the edge architecture must be implemented to achieve a realistic performance. / I HRC-scenarier mellan människor och robotar där människor och robotar arbetar tillsammans och delar samma arbetsyta finns det risk för potentiell fara som kan uppstå. I detta arbete har en AI-baserad lösning för riskanalys utvecklats för att identifiera alla tillstånd som kan skada en robot och dess miljö. Informationen från riskanalys används i en riskreduceringsmodul för att minska risken för att vara i en farlig situation. Målet är att utveckla säkerhet för HRC-scenarier med olika AI-algoritmer och att kontrollera möjligheterna att förbättra systemets effektivitet utan att kompromissa med säkerheten.Denna rapport presenterar strategier för riskreducering som byggdes ovanpå robotens styrsystem och baserade på ISO 15066-standarden. Var och en av dem använder semantisk information (scendiagram) om robotens miljö och förändrar robotens rörelse genom skalning av hastighet. Den första implementetationen av riskreducerande strategi använder Fuzzy Logic System. Detta system analyserade de mest riskabla objektens egenskaper för att justera robotens hastighet i enlighet därmed. Den andra implementeringen använder förstärkningslärande och betraktade varje objekts egenskaper. Tre nätverk (fully connected network, convolutional neural network, and hybrid network) implementeras för att uppskatta Q-värde-funktionen. Dessutom implementerade vi också lokaloch edge-arkitektur för att beräkna beräkningsprestanda på den verkliga roboten. Varje modell utvärderas genom att mäta säkerhetsaspekten och robotens prestanda i ett simulerat lagerscenario. Alla riskreduceringsmoduler kunde minska risken för potentiell fara. Fuzzy logicsystem kunde öka säkerhetsaspekten med minsta effektivitetsminskning. Förstärkningsinlärningsmodellen har säkrare drift men har en mer begränsad effektivitet än det fuzzy logiska systemet. I allmänhet fungerar fuzzy logicsystem upp till 28 % snabbare än förstärkningslärande men komprometterar upp till 23 % när det gäller säkerhet (medelrisk hastighetsvärde). När det gäller beräkningsprestanda utfördes kantberäkningen snabbare än lokal beräkning. Flaskhalsen för processen var scengrafgenerering som analyserade en bild för att producera information för säkerhetsanalys. Det tog cirka 15 sekunder att köra scengrafgenerering på robotens CPU och 0,3 sekunder på en kantenhet. Modulen för riskreducering kan väljas beroende på KPI för lagerdriften medan edge-arkitekturen måste implementeras för att uppnå en realistisk prestanda.

[pt] Os sistemas de inferência fuzzy são técnicas de aprendizado de máquina que possuem a capacidade de modelar incertezas matematicamente. Eles são divididos em sistemas de inferências fuzzy tipo-1 e fuzzy tipo-2. O sistema de inferência fuzzy tipo-1 vem sendo amplamente aplicado na solução de diversos problemas referentes ao aprendizado de máquina, tais como, controle, classificação, clusterização, previsão, dentre outros. No entanto, por apresentar uma melhor modelagem matemática das incertezas, o sistema de inferência fuzzy tipo-2 vem ganhando destaque ao longo dos anos. Esta melhora modelagem vem também acompanhada de um aumento do esforço matemático e computacional. Visando reduzir tais pontos para solucionar problemas de classificação, este trabalho apresenta o desenvolvimento e a comparação de duas funções de pertinência Gaussiana para um sistema de inferência fuzzy tipo-2 intervalar usando o método superior e inferior. São utilizadas as funções de pertinência Gaussiana com incerteza na média e com incerteza no desvio padrão. Ambos os modelos fuzzy abordados neste trabalho são treinados por algoritmos baseados em informações de primeira ordem. Além disso, este trabalho propõe a extensão dos modelos fuzzy tipo-2 intervalar para apresentarem múltiplas saídas, reduzindo significativamente o custo computacional na solução de problemas de classificação multiclasse. Finalmente, visando contextualizar a utilização desses modelos em aplicações de engenharia mecânica, este trabalho apresenta a solução de um problema de detecção de falhas em turbinas a gás, utilizadas em aeronaves. / [en] Fuzzy logic systems are machine learning techniques that can model mathematically uncertainties. They are divided into type-1 fuzzy, and type-2 fuzzy logic systems. The type-1 fuzzy logic system has been widely applied to solve several problems related to machine learning, such as control, classification, clustering, prediction, among others. However, as it presents a better mathematical modeling of uncertainties, the type-2 fuzzy logic system has received much attention over the years. This modeling improvement is also accompanied by an increase in mathematical and computational effort. Aiming to reduce these issues to solve classification problems, this work presents the development and comparison of two Gaussian membership functions for a type-2 interval fuzzy logic system using the upper and lower method. Gaussian membership functions with uncertainty in the mean and with uncertainty in the standard deviation are used. Both fuzzy models covered in this work are trained by algorithms based on first order information. Furthermore, this work proposes the extension of interval type-2 fuzzy models to present multiple outputs, significantly reducing the computational cost in solving multiclass classification problems. Finally, aiming to contextualize the use of these models in mechanical engineering applications, this work presents the solution of a problem of fault detection in aircraft gas turbines.

The major contribution of this dissertation is the development of a new integrated measurement and control system for intermixed biogas and photovoltaic systems to achieve safe and optimal energy usage. Literature and field studies show that existing control methods fall short of comprehensive system optimization and fault diagnosis, hence the need to re-look these control methods. The control strategy developed in this dissertation is a considerable enhancement on existing strategies as it incorporates intelligent fuzzy logic algorithms based on C source codes developed on the MPLABX programming environment. Measurements centered on the PIC18F4550 microcontroller were carried out on existing biogas and photovoltaic installations. The designed system was able to accurately predict digester stability, quantify biogas output and carry out biogas fault detection and control. Optimized battery charging and photovoltaic fault detection and control was also successfully implemented. The system optimizes the operation and performance of biogas and photovoltaic energy generation. / Electrical Engineering / M. Tech. (Electrical Engineering)

Composite structures are extremely useful for aerospace, automotive, marine and civil applications due to their very high specific structural properties. These structures are subjected to severe dynamic loading in their service life. Repeated exposure to these severe loading conditions can induce structural damage which ultimately may precipitate a catastrophic failure. Therefore, an interest in the continuous inspection and maintenance of engineering structures has grown tremendously in recent years. Sensitive aerospace applications can have small design margins and any inadequacy in knowledge of the system may cause design failure. Structures made from composite materials posses complicated failure mechanism as compared to those made from conventional metallic materials. In composite structural design, it is hence very important to properly model geometric intricacies and various imperfections such as delaminations and cracks. Two important issues are addressed in this thesis: (1) structural modeling of nonlinear delamination and uncertainty propagation in nonlinear characteristics of composite plate structures and (2) development of a model based damage detection system to handle uncertainty issues. An earlier proposed shear deformable C0 composite plate finite element is modified to alleviate modeling uncertainty issues associated with a damage detection problem. Parabolic variation of transverse shear stresses across the plate thickness is incorporated into the modified formulation using mixed shear interpolation technique. Validity of the proposed modification is established through available literature. Correction of the transverse shear stress term in the formulation results in about 2 percent higher solution accuracy than the earlier model. It is found that the transverse shear effect increases with higher modes of the plate deformation. Transverse shear effects are more prominent in sandwich plates. This refined composite plate finite element is used for large deformation dynamic analysis of delaminated composite plates. The inter-laminar contact at the delaminated region in composite plates is modeled with the augmented Lagrangian approach. Numerical simulations are carried out to investigate the effect of delamination on the nonlinear transient behavior of composite plates. Results obtained from these studies show that widely used unconditionally stable β-Newmark method presents numerical instability problems in the transient simulation of delaminated composite plate structures with large deformation. To overcome this instability issue, an energy and momentum conserving composite implicit time integration scheme presented by Bathe and Baig is used for the nonlinear dynamic analysis. It is also found that a proper selection of the penalty parameter is very crucial in the simulation of contact condition. It is shown that an improper selection of penalty parameter in the augmented Lagrangian formulation may lead to erroneous prediction of dynamic response of composite delaminated plates. Uncertainties associated with the mathematical characterization of a structure can lead to unreliable damage detection. Composite structures also show considerable scatter in their structural response due to large uncertainties associated with their material properties. Probabilistic analysis is carried out to estimate material uncertainty effects in the nonlinear frequencies of composite plates. Monte Carlo Simulation with Latin Hypercube Sampling technique is used to obtain the variance of linear and nonlinear natural frequencies of the plate due to randomness in its material properties. Numerical results are obtained for composite plates with different aspect ratio, stacking sequence and oscillation amplitude ratio. It is found that the nonlinear frequencies show increasing non-Gaussian probability density function with increasing amplitude of vibration and show dual peaks at high amplitude ratios. This chaotic nature of the dispersion of nonlinear eigenvalues is also revealed in eigenvalue sensitivity analysis. For fault isolation, variations in natural frequencies, modal curvatures and curvature damage factors due to damage are investigated. Effects of various physical uncertainties like, material and geometric uncertainties on the success of damage detection is studied. A robust structural damage detection system is developed based on the statistical information available from the probabilistic analysis carried out on beam type structures. A new fault isolation technique called sliding window defuzzifier is proposed to maximize the success rate of a Fuzzy Logic System (FLS) in damage detection. Using the changes in structural measurements between the damaged and undamaged state, a fuzzy system is generated and the rule-base and membership functions are generated using probabilistic informations. The FLS is demonstrated using frequency and mode shape based measurements for various beam type structures such as uniform cantilever beam, tapered beam in single as well as in multiple damage conditions. The robustness of the FLS is demonstrated with respect to the highly uncertain input information called measurement deltas (MDs). It is said, if uncertainty level is larger than or close to the changes in damage indicator due to damage, the true information would be submerged in the noise. Then the actual damaged members may not be identified accurately and/or the healthy members may be wrongly detected as damaged giving false warning. However, this being the case, the proposed FLS with new fault isolation technique tested with these noisy data having large variation and overlaps shows excellent robustness. It is observed that the FLS accurately predicts and isolates the damage levels up-to considerable uncertainty and noise levels in single as well as multiple damage conditions. The robustness of the FLS is also demonstrated for delamination detection in composite plates having very high material property uncertainty. Effects of epistemic uncertainty on damage detection in composite plates is addressed. The effectiveness of the proposed refined Reddy type shear deformable composite plate element is demonstrated for reducing the modeling or epistemic uncertainty in delamination detection.