• ETRI Journal
• /
• v.33 no.6
• /
• pp.924-934
• /
• 2011

Sensor networks play an important role in making the dream of ubiquitous computing a reality. With a variety of applications, sensor networks have the potential to influence everyone's life in the near future. However, there are a number of issues in deployment and exploitation of these networks that must be dealt with for sensor network applications to realize such potential. Localization of the sensor nodes, which is the subject of this paper, is one of the basic problems that must be solved for sensor networks to be effectively used. This paper proposes a probabilistic support vector machine (SVM)-based method to gain a fairly accurate localization of sensor nodes. As opposed to many existing methods, our method assumes almost no extra equipment on the sensor nodes. Our experiments demonstrate that the probabilistic SVM method (PSVM) provides a significant improvement over existing localization methods, particularly in sparse networks and rough environments. In addition, a post processing step for PSVM, called attractive/repulsive potential field localization, is proposed, which provides even more improvement on the accuracy of the sensor node locations.

• Journal of Electrical Engineering and information Science
• /
• v.3 no.3
• /
• pp.355-365
• /
• 1998

The purposes of this study are to propose a network for the characterizing of the input data and to show how to design predictive neural net재가 expert system which doesn't need previous knowledge base. We derived this network from the radial basis function networks(RBFN), and named it as a statistical EBFN. The proposed network can replace the statistical methods for analyzing dynamic relations between target disease and other parameters in medical studies. We compared statistical RBFN with the probabilistic neural network(PNN) and fuzzy logic(FL). And we testified our method in the diabetes prediction and compared our method with the well-known multilayer perceptron(MLP) neural network one, and showed good performance of our network. At last, we developed the diabetes prediction expert system based on the proposed statistical RBFN without previous knowledge base. Not only the applicability of the characterizing of parameters related to diabetes and construction of the diabetes prediction expert system but also wide applicabilities has the proposed statistical RBFN to other similar problems.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.28B no.10
• /
• pp.831-841
• /
• 1991

A combined model of probabilistic and MLP(multi layer perceptron) model is proposed for the pattern classification of EMG( electromyogram) signals. The MLP model has a problem of not guaranteeing the global minima of error and different quality of approximations to Bayesian probabilities. The probabilistic model is, however, closely related to the estimation error of model parameters and the fidelity of assumptions. A proper combination of these will reduce the effects of the problems and be robust to input variations. Proposed model is able to get the MAP(maximum a posteriori probability) in the probabilistic model by estimating a priori probability distribution using the MLP model adaptively. This method minimize the error probability of the probabilistic model as long as the realization of the MLP model is optimal, and this is a good combination of the probabilistic model and the MLP model for the usage of MLP model reliability. Simulation results show the benefit of the proposed model compared to use the Mlp and the probabilistic model seperately and the average calculation time fro classification is about 50ms in the case of combined motion using an IBM PC 25 MHz 386model.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1997.10a
• /
• pp.181-184
• /
• 1997

Practical applications of learning systems usually involve complex domains exhibiting nonlinear behavior and dilution by noise. Consequently, an intelligent system must be able to adapt to nonlinear processes as well as probabilistic phenomena. An important class of application for a knowledge based systems in prediction: forecasting the future trajectory of a process as well as the consequences of any decision made by e system. This paper examines the robustness of data mining tools under varying levels of noise while predicting nonlinear processes in the form of chaotic behavior. The evaluated models include the perceptron neural network using backpropagation (BPN), the recurrent neural network (RNN) and case based reasoning (CBR). The concepts are crystallized through a case study in predicting a Henon process in the presence of various patterns of noise.

• ETRI Journal
• /
• v.31 no.2
• /
• pp.209-214
• /
• 2009

In this paper, we propose a novel test methodology for the detection of catastrophic and parametric faults present in analog very large scale integration circuits. An automatic test pattern generation algorithm is proposed to generate piece-wise linear (PWL) stimulus using wavelets and a genetic algorithm. The PWL stimulus generated by the test algorithm is used as a test stimulus to the circuit under test. Faults are injected to the circuit under test and the wavelet coefficients obtained from the output response of the circuit. These coefficients are used to train the neural network for fault detection. The proposed method is validated with two IEEE benchmark circuits, namely, an operational amplifier and a state variable filter. This method gives 100% fault coverage for both catastrophic and parametric faults in these circuits.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.15 no.2
• /
• pp.395-406
• /
• 1995

Classification of flaws in weldments from their ultrasonic scattering signals is very important in quantitative nondestructive evaluation. This problem is ideally suited to a modern ultrasonic pattern recognition technique. Here brief discussion on systematic approach to this methodology is presented including ultrasonic feature extraction, feature selection and classification. A stronger emphasis is placed on probabilistic neural networks as efficient classifiers for many practical classification problems. In an example probabilistic neural networks are applied to classify flaws in weldments into 3 classes such as cracks, porosity and slag inclusions. Probabilistic nets are shown to be able to exhibit high performance of other classifiers without any training time overhead. In addition, forward selection scheme for sensitive features is addressed to enhance network performance.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.2011-2016
• /
• 2004

Study of human-robot communication is one of the most important research areas. Among various communication media, any useful law we find in voice communication in human-human interactions, is significant in human-robot interactions too. Control strategy of most of such systems available at present is on/off control. These robots activate a function if particular word or phrase associated with that function can be recognized in the user utterance. Recently, there have been some researches on controlling robots using information rich fuzzy commands such as "go little slowly". However, in those works, although the voice command interpretation has been considered, learning from such commands has not been treated. In this paper, learning from such information rich voice commands for controlling a robot is studied. New concepts of the coach-player model and the sub-coach are proposed and such concepts are also demonstrated for a PA-10 redundant manipulator.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.21 no.4
• /
• pp.475-480
• /
• 2011

In this paper, we propose a collision risk assesment system. First, using Kalman Filter, we estimate the information of pedestrian, and second, we compute the collision probability using Monte Carlo Simulations(MCS) and neural network(NN). And we update the collision risk using time history which is called belief. Belief update consider not only output of Kalman Filter of only current time step but also output of Kalman Filter up to the first time step to current time step. The computer simulations will be shown the validity of our proposed method.

• The Journal of the Acoustical Society of Korea
• /
• v.17 no.4E
• /
• pp.3-10
• /
• 1998

This paper concerns the dynamical behavior, in probabilistic sense, of a feedforward neural network performing auto association for novelty. Networks of retinotopic topology having a one-to-one correspondence between and output units can be readily trained using back-propagation algorithm, to perform autoassociative mappings. A novelty filter is obtained by subtracting the network output from the input vector. Then the presentation of a "familiar" pattern tends to evoke a null response ; but any anomalous component is enhanced. Such a behavior exhibits a promising feature for enhancement of weak signals in additive noise. As an analysis of the novelty filtering, this paper shows that the probability density function of the weigh converges to Gaussian when the input time series is statistically characterized by nonsymmetrical probability density functions. After output units are locally linearized, the recursive relation for updating the weight of the neural network is converted into a first-order random differential equation. Based on this equation it is shown that the probability density function of the weight satisfies the Fokker-Planck equation. By solving the Fokker-Planck equation, it is found that the weight is Gaussian distributed with time dependent mean and variance.

• The Journal of the Acoustical Society of Korea
• /
• v.35 no.6
• /
• pp.501-509
• /
• 2016

We propose a method to improve the robustness of speaker verification on short test utterances. The accuracy of the state-of-the-art i-vector/probabilistic linear discriminant analysis systems can be degraded when testing utterance durations are short. The proposed method compensates for utterance variations of short test feature vectors using deep neural networks. We design three different types of DNN (Deep Neural Network) structures which are trained with different target output vectors. Each DNN is trained to minimize the discrepancy between the feed-forwarded output of a given short utterance feature and its original long utterance feature. We use short 2-10 s condition of the NIST (National Institute of Standards Technology, U.S.) 2008 SRE (Speaker Recognition Evaluation) corpus to evaluate the method. The experimental results show that the proposed method reduces the minimum detection cost relative to the baseline system.