• ETRI Journal
• /
• 제45권4호
• /
• pp.557-569
• /
• 2023

Traditional deterministic channel modeling is accurate in prediction, but due to its complexity, improving computational efficiency remains a challenge. In an alternative approach, we investigated a multilayer artificial neural network (ANN) to predict large-scale and small-scale channel characteristics in metro tunnels. Simulated high-precision training datasets were obtained by combining measurement campaign with a ray tracing (RT) method in a metro tunnel. Performance on the training data was used to determine the number of hidden layers and neurons of the multilayer ANN. The proposed multilayer ANN performed efficiently (10 s for training; 0.19 ms for prediction), and accurately, with better approximation of the RT data than the single-layer ANN. The root mean square errors (RMSE) of path loss (2.82 dB), root mean square delay spread (0.61 ns), azimuth angle spread (3.06°), and elevation angle spread (1.22°) were impressive. These results demonstrate the superior computing efficiency and model complexity of ANNs.

• Nuclear Engineering and Technology
• /
• 제48권1호
• /
• pp.64-71
• /
• 2016

Void fraction is an important parameter in the oil industry. This quantity is necessary for volume rate measurement in multiphase flows. In this study, the void fraction percentage was estimated precisely, independent of the flow regime in gas-liquid two-phase flows by using ${\gamma}-ray$ attenuation and a multilayer perceptron neural network. In all previous studies that implemented a multibeam ${\gamma}-ray$ attenuation technique to determine void fraction independent of the flow regime in two-phase flows, three or more detectors were used while in this study just two NaI detectors were used. Using fewer detectors is of advantage in industrial nuclear gauges because of reduced expense and improved simplicity. In this work, an artificial neural network is also implemented to predict the void fraction percentage independent of the flow regime. To do this, a multilayer perceptron neural network is used for developing the artificial neural network model in MATLAB. The required data for training and testing the network in three different regimes (annular, stratified, and bubbly) were obtained using an experimental setup. Using the technique developed in this work, void fraction percentages were predicted with mean relative error of <1.4%.

• Transactions on Control, Automation and Systems Engineering
• /
• 제2권3호
• /
• pp.169-174
• /
• 2000

In this paper, Haar wavelet-based neural network is described for the identification and control of discrete-time nonlinear dynamical systems. Wavelets are suited to depict functions with local nonlinearities and fast variations because of their intrinsic properties of finite support and self-similarity. Due to the orthonormal properties of Haar wavelet functions, wavelet neural networks result in a greatly simplified training problem. This wavelet-based scheme performs adaptively both the identification of nonlinear functions and the control of the overall system, while the multilayer neural network is applied to the control system just after its sufficient learning of the unknown functions. Simulation shows that the wavelet network can be a good alternative to a multilayer neural network with backpropagation.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1998년도 추계종합학술대회 논문집
• /
• pp.1285-1288
• /
• 1998

In this paper, we propose a new learning algorithm for multilayer neural networks. In the error backpropagation that is widely used for training multilayer neural networks, weights are adjusted to reduce the error function that is sum of squared error for all the neurons in the output layer of the network. In the proposed learning algorithm, we consider each output of the output layer as a function of weights and adjust the weights directly so that the output neurons produce the desired outputs. Experiments show that the proposed algorithm outperforms the backpropagation learning algorithm.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1996년도 한국자동제어학술회의논문집(국내학술편); 포항공과대학교, 포항; 24-26 Oct. 1996
• /
• pp.1173-1176
• /
• 1996

Multilayer feedforward networks may be applied to identify the deterministic relationship between input and output data. When the results from the network require a high level of assurance, considering of the stochastic relationship between the data may be very important. The variance is one of the useful parameters to represent the stochastic relationship. This paper presents a new algorithm for a multilayer feedforward network to learn the variance of dispersed data without preliminary calculation of variance. In this paper, the network with this learning algorithm is named as a variance learning neural network(VALEAN). Computer simulation examples are utilized for the demonstration and the evaluation of VALEAN.

• 전자공학회논문지B
• /
• 제33B권12호
• /
• pp.95-105
• /
• 1996

This paper presents structure optimization of a feedforward neural netowrk controller using the genetic algorithm. It is important to design the neural network with minimum structure for fast response and learning. To minimize the structure of the feedforward neural network, a genralization of multilayer neural netowrks, the genetic algorithm uses binary coding for the structure and floating-point coding for weights. Local search with an on-line learnign algorithm enhances the search performance and reduce the time for global search of the genetic algorithm. The relative fitness defined as the multiplication of the error and node functions prevents from premature convergence. The feedforward neural controller of smaller size outperformed conventional multilayer perceptron network controller.

• 전자공학회논문지B
• /
• 제32B권12호
• /
• pp.1670-1679
• /
• 1995

In this paper, an optimal variable structure controller with a multilayer neural inverse identifier is proposed. A multilayer neural network with error back propagation learning algorithm is used for construction the neural inverse identifier which is an observer of the external disturbances and the parameter variations of the system. The variable structure controller with the multilayer neural inverse identifier not only needs a small part of a priori knowledge of the bounds of external disturbances and parameter variations but also alleviates the chattering magnitude of the control input. Also, an optimal sliding line is designed by the optimal linear regulator technique and an integrator is introduced for solving the reaching phase problem. Computer simulation results show that the proposed approach gives the effective control results by reducing the chattering magnitude of control input.

• 한국정밀공학회지
• /
• 제14권6호
• /
• pp.121-127
• /
• 1997

Multilayer feedforward networks may be applied to identify the deterministic relationship between input and output data. When the results from the network require a high level of assurance, consideration of the stochastic relationship between the input and output data may be very important. Variance is one of the effective parameters to deal with the stochastic relationship. This paper presents a new algroithm for a multilayer feedforward network to learn the variance of dispersed data without preliminary calculation of variance. In this paper, the network with this learning algorithm is named as a variance learning neural network(VALEAN). Computer simulation examples are utilized for the demonstration and the evaluation of VALEAN.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2000년도 제15차 학술회의논문집
• /
• pp.394-394
• /
• 2000

This Paper proposes a new nonlinear partial least square method that extends the linear PLS. Proposed nonlinear PLS uses self-organizing feature map as PLS outer relation and multilayer neural network as PLS inner regression method.

• 대한전기학회논문지
• /
• 제40권8호
• /
• pp.782-791
• /
• 1991

In the fields of pattern recognition such as speech recognition, several new techniques using Artifical Neural network Models have been proposed and implemented. In particular, the Multilayer Perception Model has been shown to be effective in static speech pattern recognition. But speech has dynamic or temporal characteristics and the most important point in implementing speech recognition systems using Artificial Neural Network Models for continuous speech is the learning of dynamic characteristics and the distributed cues and contextual effects that result from temporal characteristics. But Recurrent Multilayer Perceptron Model is known to be able to learn sequence of pattern. In this paper, the results of applying the Recurrent Model which has possibilities of learning tedmporal characteristics of speech to phoneme recognition is presented. The test data consist of 144 Vowel+ Consonant + Vowel speech chains made up of 4 Korean monothongs and 9 Korean plosive consonants. The input parameters of Artificial Neural Network model used are the FFT coefficients, residual error and zero crossing rates. The Baseline model showed a recognition rate of 91% for volwels and 71% for plosive consonants of one male speaker. We obtained better recognition rates from various other experiments compared to the existing multilayer perceptron model, thus showed the recurrent model to be better suited to speech recognition. And the possibility of using Recurrent Models for speech recognition was experimented by changing the configuration of this baseline model.