Notification:
We are currently experiencing intermittent issues impacting performance. We apologize for the inconvenience.
By Topic

Computational Intelligence for Measurement Systems and Applications, 2008. CIMSA 2008. 2008 IEEE International Conference on

Date 14-16 July 2008

Filter Results

Displaying Results 1 - 25 of 32
  • CIMSA 2008 - IEEE conference on computational intelligence for measurement systems and applications proceedings

    Publication Year: 2008 , Page(s): c1
    Save to Project icon | Request Permissions | PDF file iconPDF (183 KB)  
    Freely Available from IEEE
  • [Copyright notice]

    Publication Year: 2008 , Page(s): ii - iii
    Save to Project icon | Request Permissions | PDF file iconPDF (193 KB)  
    Freely Available from IEEE
  • Table of contents

    Publication Year: 2008 , Page(s): iv - vii
    Save to Project icon | Request Permissions | PDF file iconPDF (131 KB)  
    Freely Available from IEEE
  • Message from the Chairs

    Publication Year: 2008 , Page(s): viii - ix
    Save to Project icon | Request Permissions | PDF file iconPDF (127 KB)  
    Freely Available from IEEE
  • Uncertainty mesurement in video and infrared cameras system

    Publication Year: 2008 , Page(s): 1 - 3
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (370 KB) |  | HTML iconHTML  

    Video and infrared cameras can be used in a variety of applications for measurements and recognitions. But these entire situations are affected by different degrees of uncertainty. In order to evaluate this uncertainty we propose a new method of mesurement based on fuzzy similarity. From two independent systems based on video and respectively on infrared cameras, two uncertainty evaluations are obtained. Based on these data, an information fusion is performed, taking in consideration the complementary behavior in time of the systems. The final result is proved to outperform each of the independently taken uncertainty measurements. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural technologies for increasing the GPS position accuracy

    Publication Year: 2008 , Page(s): 4 - 8
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (682 KB) |  | HTML iconHTML  

    Aim of this paper is to present a method to improve the accuracy of a GPS receiver. It is well known that there are many factors affecting the accuracy of a GPS receiver. In this work, the authors point out that many of these factors, considered in a given geographic area, have a certain periodicity. An important example of this kind of factors is the sky satellite position relative to receiver. The proposed method uses a neural network to correct the position computed by the receiver. The neural network is trained to learn the errors introduced into the measuring system by the cyclic phenomenon in the various hours of the day. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Coupling Kernel Principal Component Analysis with ANN for improving analysis accuracy of seven-component alkane gaseous mixture

    Publication Year: 2008 , Page(s): 9 - 13
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (542 KB) |  | HTML iconHTML  

    To further improving the analysis accuracy of Artificial Neural Networks (ANN) model for quantitative analysis of seven-component alkane gaseous mixtures composed of methane, ethane, propane, isobutane, n-butane, isopentane, and n-pentane, the Kernel Principal Component Analysis (KPCA) technique was proposed to couple with it. The gaseous mixtures were measured by a novel Acousto-Optic Tunable Filter Near Infrared (AOTF-NIR) spectrometer. KPCA mapped the NIR spectral data of gaseous mixtures by a Gaussian kernel to a high-dimensional feature space and implemented feature extraction in it. As input variables, the extracted features were fed into a three-layered ANN to create quantitative analysis model of above-mentioned seven component gases. The performance of KPCA-NN model was assessed by Root Mean Square Error of Prediction (RMSEP) of testing set. The RMSEP of seven components by KPCA-ANN were less than 0.361%. Comparing with the ANN model without KPCA feature extraction, the KPCA-ANN model obtained the less RMSEP values. The research results indicated that the KPCA-NN model shows higher analysis accuracy than ANN model. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A quality of performance model for evaluating post-stroke patients

    Publication Year: 2008 , Page(s): 14 - 18
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (539 KB) |  | HTML iconHTML  

    Augmented reality (AR) has recently emerged as an assistive tool for effective diagnosis and rehabilitation intervention. However, measuring the quality of performance (QoP) of patients has gained limited attention from the research community. The objective of this paper is to propose and test a evaluation taxonomy for an AR-based stroke patient rehabilitation system that is currently under development at the MCRlab, University of Ottawa. The taxonomy is modeled using a fuzzy logic inference (FLI) system to quantitatively measure the QoP of the patient and eventually provide the therapist with discrete recommendation regarding the progress of patient treatment. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A DCT based nonlinear predictive coding for feature extraction in speech recognition systems

    Publication Year: 2008 , Page(s): 19 - 22
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (442 KB) |  | HTML iconHTML  

    Speech representation strategies play a key role in automatic speech recognition systems. In this study, a nonlinear procedure has been proposed to overcome the complexities of speech sequence representations. The proposed method may be considered as an extension of nonlinear predictive coding representation procedure in cosine transform domain. The best results belong to classification of nonlinear behaved stop phonemes (i.e. /b/, /d/, /g/) in TIMIT database which show good performance while reducing the computational complexity in comparison to standard NPC. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Radial basis function networks with quantized parameters

    Publication Year: 2008 , Page(s): 23 - 27
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (636 KB) |  | HTML iconHTML  

    A RBFN implemented with quantized parameters is proposed and the relative or limited approximation property is presented. Simulation results for sinusoidal function approximation with various quantization levels are shown. The results indicate that the network presents good approximation capability even with severe quantization. The parameter quantization decreases the memory size and circuit complexity required to store the network parameters leading to compact mixed-signal circuits proper for low-power applications. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Image processing for granulometry analysis via neural networks

    Publication Year: 2008 , Page(s): 28 - 32
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (881 KB) |  | HTML iconHTML  

    The analysis of granulometry of substances is relevant in a great variety of the research and industrial applications as such as the pharmaceutical sector, the food sector, the basic materials production and in the concrete and wood panel industries. This analysis is important since many relevant properties of the materials can depend on the distribution of the particles sizes/shapes during the production. In this work we present an innovative method capable to estimate the particles size distribution in an image without the use of segmentation techniques by using neural networks. The paper contribution is twofold. The proposed method presents a set of techniques based on wavelet analysis and image processing techniques suitable to extract relevant features for the granulometry analysis. Then, the extracted set of features is used as input to neural networks in order to achieve the classification of each single pixel accordingly to the probability to belong to a specific class of particles size (a single band in the histogram of the distribution of the particles size). The produced outputs have been used to perform the estimation of the particle granulometry contained in the image. Results are encouraging and show the effectiveness of the proposed method. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multiscale windowed denoising and segmentation of hyperspectral images

    Publication Year: 2008 , Page(s): 33 - 37
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (551 KB) |  | HTML iconHTML  

    This paper presents the effects of multiscale windowed denoising of spectral signatures before segmentation of hyperspectral images. In the proposed denoising approach it is intended to exploit both spectral and spatial information of the hyperspectral images by using wavelets and principal component analysis. The windowed structure incorporated for this method exploits spatial information by making use of possibly highly correlated pixels. In addition to the proposed method, the segmented PCA is also investigated and compared in the experimental results with a proper modification. In the segmentation process, the K-means and fuzzy-ART algorithms are used. Especially fuzzy-ART is a fast learning network and can be used in high dimensional and high volume data such as hyperspectral images. In the experiments it has been shown that multiscale windowed principal component denoising has positive effects on the segmentation/clustering level. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Improvements to the bacterial memetic algorithm used for fuzzy rule base extraction

    Publication Year: 2008 , Page(s): 38 - 43
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (466 KB) |  | HTML iconHTML  

    In this paper we discuss new methods to improve the bacterial memetic algorithm (BMA) used for fuzzy rule base extraction. The first two methods are knot order violation handling methods which improves the performance of the BMA rather in the case of more complex fuzzy rule base. The third method is a new modification of the BMA in which the order of the operators is modified. This method improves the performance of the BMA rather in the case of less complex fuzzy rule base. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fuzzy flip-flop based neural network as a function approximator

    Publication Year: 2008 , Page(s): 44 - 49
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1656 KB) |  | HTML iconHTML  

    Artificial neural networks and fuzzy logic systems, in the context of approximate reasoning, share common features and techniques. A family of fuzzy flip-flops is proposed, based on an artificial neural network-like structure which is suitable for approximating many-input one-output nonlinear functions. The neurons in the multilayer perceptron networks typically employ sigmoidal activation functions. The next state of the fuzzy J-K flip-flops (F3) using Yager and Dombi operators present quasi-S-shaped characteristics. The paper proposes the investigation of the possibility of constructing multilayer perceptrons from such fuzzy units. Each of the two candidates for F3-based neurons is examined for its training capability by evaluating and comparing the approximation properties in the context of different transcendental functions with one-input and multi-input cases. Simulation results are presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Toward a theory of validation of hybrid MinMax FuzzyNeuro systems

    Publication Year: 2008 , Page(s): 50 - 53
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (438 KB) |  | HTML iconHTML  

    The validation and verification (V&V) of hybrid fuzzyneuro (HFN) or hybrid neurofuzzy (HNF) systems becomes of increasing concern as these systems are fielded and embedded in the every day operations of medical diagnosis, pattern recognition, fuzzy control and other industries-particularly so when life-critical and environment-critical aspects are involved. We provide in this paper a V&V perspective on the nature of HFN components, an appropriate life-cycle, and applicable systematic formal testing approaches. We consider why HFN V&V may be both easier and harder than traditional means, and we conclude with a series of practical V&V guidelines. Validation of HFN systems brings us to a systematic study of value approximation performed during the inference phase. It is accepted that generalization capability is proportional to value approximation. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Analysis of the effects of bearings-only sensors on the performance of the neural extended kalman filter tracking system

    Publication Year: 2008 , Page(s): 54 - 59
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (477 KB) |  | HTML iconHTML  

    The neural extended Kalman filter (NEKF) has proven to be a quality maneuver target tracking system when the sensors provide a fully observable measurement, such as a radarpsilas range-bearing measurement or a position report. As with any state estimation technique, the NEKF requires observability in order to estimate the target track states. Observability is needed as well to train the weights of the neural network, since the neural network training paradigm is coupled to the target states. Passive sensor systems, such as electronic surveillance measures and passive sonar arrays, provide an angle-only measurement. Such bearings-only measurements make the tracking system an unobservable system. For a Kalman filter estimator, this will result in the eigenvalues of the error covariance matrix to grow without bound. For the NEKF, since both the target state and the weights of the neural network are affected by the lack of observability, the results could be more pronounced. In this paper, the application of the NEKF in bearings-only tracking problems is analyzed to determine the effects on performance. The analyzed cases look at a single sensor platform in four important scenarios: a stationary platform and straight-line target, a stationary platform and a maneuvering target, a maneuvering platform and a straight-line target, and a maneuvering platform and a maneuvering target. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Neural models for ambient temperature modelling

    Publication Year: 2008 , Page(s): 60 - 63
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (496 KB) |  | HTML iconHTML  

    In this work we show how to model ambient temperature through neural models. In particular we tried feed forward and fully recurrent architectures, trained with the back-propagation and evolutionary algorithms, to estimate the monthly average temperature and compared the results to the nearest neighbor approach. Therefore, the best neural model has been tested to get hourly estimations. We compared the outcomes to a well known tool which doesn't have such an estimation capability and results show that the proposed approach clearly outperforms the traditional ones. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Application of data approximation and classification in measurement systems - comparison of “neural network” and “Least Squares” approximation

    Publication Year: 2008 , Page(s): 64 - 69
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (466 KB) |  | HTML iconHTML  

    In measurement systems, environmental conditions are measured based on predefined scenarios. Measured data are then processed in either a decentralized or centralized manner. In advanced systems (especially for distributed data processing), taking artificial intelligence features into consideration could improve measurement performance and reliability. It is assumed as autonomy in measurement system which leads to distributed ldquointelligent data measurement and processingrdquo. In this paper, two different methodologies for ldquotemperature predictionrdquo are compared. A discussion concerning the classification of recorded data is then presented. Both a mathematical approach, the so-called ldquoleast squaresrdquo approach, and a model-free approach, called back-propagation, are applied and compared for temperature approximation. After approximation, the predicted temperature values are compared with real temperature records for classification purposes. The ldquoclassification mechanismrdquo includes signal processing features for improving performance. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Urban pollution monitoring through opportunistic mobile sensor networks based on public transport

    Publication Year: 2008 , Page(s): 70 - 74
    Cited by:  Papers (5)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1609 KB) |  | HTML iconHTML  

    The development of an opportunistic sensor network deployed on regular public transport vehicles with the aim of obtaining a flexible pollution monitoring system over large urban areas is presented. Georeferenced pollution data is acquired by a modular autonomous sensing system placed on vehicles which has been developed and is being currently tested. Short and long range communication systems are used to transmit data from the mobile sources to the central data processing and mapping unit. Within this unit an application to represent the geopositioned pollutant measurements has been implemented based on Google Earth. This provides the user with an interface allowing the study of the evolution of the gas concentrations along a given bus route as well as on the whole urban area. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Developing a neural network model for magnetic yoke structure

    Publication Year: 2008 , Page(s): 75 - 78
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (448 KB) |  | HTML iconHTML  

    Magnetic flux leakage technique is used extensively to detect and characterize defects in natural gas and oil transmission pipelines. The amount of magnetic flux introduced into the test sample is an important factor in the resolution of flaw detection. It depends on the power of permanent magnets and the geometrical design of the magnetic yoke. Finite element method (FEM) is the most widely used method of analyzing magnetic yoke due to its power, accuracy and straightforwardness. On the other hand its calculations are so complicated and time consuming, and every single modification in the parameters of the problem requires a new run. In this paper, we present an innovative method to overcome the problem of heavy calculations. In this method an artificial neural network (ANN) is trained to simulate the behavior of the magnetic yoke for different design parameters with an acceptable error. Afterwards the trained ANN calculates the desired output (usually generated flux) for a new design of the yoke by generalization of the already seen samples. This new method has got two advantages over the traditional FEM. First it is very fast and second it is flexible due to modifications in parameters. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Application of reinforcement learning to improve control performance of plant

    Publication Year: 2008 , Page(s): 79 - 82
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (466 KB)  

    This paper is concerned with the development of an online Reinforcement Learning (RL) technique that significantly improves the control systems behavior. The reinforcement learner is based on Q-learning and the final controller is an artificial neural network whose weights are tuned by on line learning. In order to speed up the learning processes and prevent the plant from the instability, initially a PID is utilized as an augmented controller until the reinforcement learning becomes capable of keep the system stable and prevent the system from undesirable behavior. Example of use is presented and the effectiveness of the proposed approach is shown. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Categorization of power quality transients using higher-order statistics and competitive layers-based neural networks

    Publication Year: 2008 , Page(s): 83 - 86
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (352 KB) |  | HTML iconHTML  

    This paper deals with power-quality (PQ) event detection, classification and characterization using higher-order sliding cumulants (which are calculated over high-pass filtered signals to avoid the low-frequency 50-Hz sinusoid), whose maxima and minima are the coordinates of two-dimensional feature vectors. The classification strategy is based in competitive layers. We focus on the problem of differentiating two types of transients: short-duration (impulsive transients) and long-duration (oscillatory transients). The results show that the measured vectors are classified into clearly differentiated clusters in the feature space. The experience aims to set the foundations of an automatic procedure for PQ event detection. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Diagnosis and monitoring of complex industrial processes based on self-organizing maps and watershed transformations

    Publication Year: 2008 , Page(s): 87 - 92
    Cited by:  Papers (2)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (8838 KB) |  | HTML iconHTML  

    A cost-effective operation of complex automation systems requires the continuous diagnosis of the asset functionality. The early detection of potential failures and malfunctions, the identification and localization of present or impending component failures and, in particular, the monitoring of the underlying physical process are of crucial importance for the efficient operation of complex process industry assets. With respect to these suppositions a software agent based diagnosis and monitoring concept has been developed, which allows an integrated and continuous diagnosis of the communication network and the underlying physical process behavior. The present paper outlines the architecture of the developed distributed diagnostic concept based on software agents and presents the functionality for the diagnosis of the unknown process behaviour of the underlying automation system based on machine learning methods. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Artificial immune inspired fault detection algorithm based on fuzzy clustering and genetic algorithm methods

    Publication Year: 2008 , Page(s): 93 - 98
    Cited by:  Papers (5)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (545 KB) |  | HTML iconHTML  

    Early detection and diagnosis of incipient faults are desired for online condition monitoring and improved operational efficiency of induction motors. In this study, an artificial immune inspired fault detection algorithm based on fuzzy clustering and genetic algorithm is developed to detect broken rotor bar and broken connector faults in induction motors. The proposed algorithm uses only one phase stator current as input without the need for any other signals. The new feature signal called envelop is obtained by using Hilbert transform. This signal is examined in a phase space that is constructed by nonlinear time series analysis method. The artificial immune algorithm called negative selection is used to detect faults. The cluster centers of healthy motor phase space are obtained by fuzzy clustering method and they are taken as self patterns. The detectors of negative selection are generated by genetic algorithm. Self patterns generated by fuzzy clustering speed up the training stage of our algorithm and only small numbers of detectors are sufficient to detect any faults of induction motor. Results have demonstrated that the proposed system is able to detect faults in a three phase induction motor, successfully. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Towards artificial intelligence based automatic adaptive response analyzer for high frequency analog BIST

    Publication Year: 2008 , Page(s): 99 - 104
    Cited by:  Papers (1)
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (522 KB) |  | HTML iconHTML  

    In this paper we analyze the feasibility of a novel neural networks (NN) -based embedded self-test framework for analog devices and systems. The solution that we propose avoids signal quantization, directly dealing with original analog signals, which enables high-accuracy fault detection through lossless signal processing. This is only possible when the self-test unit is also built using analog components and works accordingly to the principles of analog computer. We use, however, powerful apparatus of discrete-time NN to find parameters of the self-test unit that would resemble the behavior of this NN. We demonstrate the efficiency of our approach using complex non-periodic non-linear analog signal. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.