• ETRI Journal
• /
• v.15 no.2
• /
• pp.35-51
• /
• 1993

This paper presents function approximation based on nonparametric estimation. As an estimation model of function approximation, a three layered network composed of input, hidden and output layers is considered. The input and output layers have linear activation units while the hidden layer has nonlinear activation units or kernel functions which have the characteristics of bounds and locality. Using this type of network, a many-to-one function is synthesized over the domain of the input space by a number of kernel functions. In this network, we have to estimate the necessary number of kernel functions as well as the parameters associated with kernel functions. For this purpose, a new method of parameter estimation in which linear learning rule is applied between hidden and output layers while nonlinear (piecewise-linear) learning rule is applied between input and hidden layers, is considered. The linear learning rule updates the output weights between hidden and output layers based on the Linear Minimization of Mean Square Error (LMMSE) sense in the space of kernel functions while the nonlinear learning rule updates the parameters of kernel functions based on the gradient of the actual output of network with respect to the parameters (especially, the shape) of kernel functions. This approach of parameter adaptation provides near optimal values of the parameters associated with kernel functions in the sense of minimizing mean square error. As a result, the suggested nonparametric estimation provides an efficient way of function approximation from the view point of the number of kernel functions as well as learning speed.

• Proceedings of the Optical Society of Korea Conference
• /
• 1991.06a
• /
• pp.35-40
• /
• 1991

In this paper, a new adaptive learning algorithm, Training by Adaptive Gain (TAG) for optical implementation of large-sized neural networks has been developed and its electro-optical implementation for 2-dimensional input and output neurons has been demostrated. The 4-dimensional global fixed interconnections and 2-dimensional adaptive gain-controls are implemented by multi-facet computer generated holograms and LCTV spatial light modulators, respectively. When the input signals pass through optical system to the output classifying layer, the TAG adaptive learning algorithm is implemented by a personal computer. The system classifies three 5$\times$5 input patterns correctly.

• Proceedings of the KIEE Conference
• /
• 2002.07d
• /
• pp.2216-2218
• /
• 2002

A system identification is to measure the output in the presence of a adequate input for the controlled system and to estimate the mathematical model in the basic of input output data. In the system identification, it is possible to estimate the true parameter values by the adjusted least squares method in the input-output case of no observed noise, and it is possible to estimate the true parameter values by the total least squares method in the input-output case with the observed noise. In recent the adjusted least squares method is suggested as a consistent estimation method in the system identification not with the observed noise input but with the observed noise output. In this paper we have developed the adjusted least squares method from the least squares method and have made certain of the efficiency in comparing the estimating results with the generating data by the computer simulations.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.43 no.6 s.312
• /
• pp.65-84
• /
• 2006

In this paper, we present an improved image formation model and propose a color image enhancement based on the model. In the presented image formation model, an input image is represented as a product of global illumination, local illumination, and reflectance. In the proposed color image enhancement, an input RGB color image is converted into an HSV color image. Under the assumption of white-light illumination, the H and S component images are remained as they are and the V component image only is enhanced based on the image formation model. The global illumination is estimated by applying a linear LPF with wide support region to the input V component image and the local illumination by applying a JND (just noticeable difference)-based nonlinear LPF with narrow support region to the processed image, where the estimated global illumination is eliminated from the input V component image. The reflectance is estimated by dividing the input V component image by the estimated global and local illuminations. After performing the gamma correction on the three estimated components, the output V component image is obtained from their product. Histogram modeling is next executed such that the final output V component image is obtained. Finally an output RGB color image is obtained from the H and S component images of the input color image and the final output V component image. Experimental results for the test image DB built with color images downloaded from NASA homepage and MPEG-7 CCD color images show that the proposed method gives output color images of very well-increased global and local contrast without halo effect and color shift.

• 제어로봇시스템학회:학술대회논문집
• /
• 2003.10a
• /
• pp.948-950
• /
• 2003

In this literature, the CMOS current mode fout quadrant differentiator circuit is proposed. The implementation is base on an appropriate input stage that converts the input current into a compressed voltage at the input capacitor ($C_{gs}$) of the CMOS driver circuit. This input voltage use as the control output current which flow to the output node by passing through a MOS active load and use it as the feedback voltage to the input node. Simulation results with level 49 CMOS model of MOSIS are given to demonstrate the correct operation of the proposed configuration. But the gain of the circuit is too low so the output differentiate current also low. The proposed differentiator is expected to find several applications in analog signal processing system.

• 제어로봇시스템학회:학술대회논문집
• /
• 1991.10a
• /
• pp.687-692
• /
• 1991

To reduce the complexity in constructing a fuzzy logic controller of multivariable systems, three major methods are presented. One is the method of constructing single-input-single-output fuzzy logic controllers after decoupling the target system. Another is the method of using fuzzy relation matrices which indicate the relation between each input and output. The other is the method of using the hierarchically classified inputs which dominantly influence one output than other inputs. Using the last two methods, simulation results of fuzzy logic controller implemented on 160MW boiler-turbine plant model are also shown.

• Proceedings of the KIPE Conference
• /
• 1996.06a
• /
• pp.85-88
• /
• 1996

New power factor correction(PFC) technique based on the averaged model of boost converter is proposed. Without measurement of input current, power factor correction scheme derived from the averaged model is presented. With the measurements of input voltage and output voltage, the control signal is generated to make the shape of the line current same as the input voltage. The characteristics of input line current distortion is analyzed by considering the generation of duty cycle.

• Transactions of the Korean Society of Mechanical Engineers B
• /
• v.25 no.4
• /
• pp.467-474
• /
• 2001

The mathematical model of a air-conditioning system is generally very complex and difficult to apply to controller design. In this paper, simple models applicable to the controller design are obtained by modeling the air-conditioning system by single-input single-output between compressor speed and indoor temperature, and by multi-input single-output between compressor speed, indoor fan speed and indoor temperature. Using these empirical models, model predictive control(MPC) technique was implemented for indoor temperature control of the air-conditioning system. It has been shown from various experiments that the indoor temperature control based on the MPC scheme yields reasonably good tracking performance with smooth changes in plant inputs. this multi-input multi-output MPC approach can be extended to multi air- conditioning systems where the conventional PID control scheme is very difficult to apply.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.12
• /
• pp.3330-3344
• /
• 2023

This paper proposes an antenna performance prediction model in the autonomous driving radar manufacturing process. Our research work is based upon a challenge dataset, Driving Radar Manufacturing Process Dataset, and a typical AutoML machine learning workflow engine, Pycaret open-source Python library. Note that the dataset contains the total 70 data-items, out of which 54 used as input features and 16 used as output features, and the dataset is properly built into resolving the multi-output regression problem. During the data regression analysis and preprocessing phase, we identified several input features having similar correlations and so detached some of those input features, which may become a serious cause of the multicollinearity problem that affect the overall model performance. In the training phase, we train each of output-feature regression models by using the AutoML approach. Next, we selected the top 5 models showing the higher performances in the AutoML result reports and applied the ensemble method so as for the selected models' performances to be improved. In performing the experimental performance evaluation of the regression prediction model, we particularly used two metrics, MAE and RMSE, and the results of which were 0.6928 and 1.2065, respectively. Additionally, we carried out a series of experiments to verify the proposed model's performance by comparing with other existing models' performances. In conclusion, we enhance accuracy for safer autonomous vehicles, reduces manufacturing costs through AutoML-Pycaret and machine learning ensembled model, and prevents the production of faulty radar systems, conserving resources. Ultimately, the proposed model holds significant promise not only for antenna performance but also for improving manufacturing quality and advancing radar systems in autonomous vehicles.

• Journal of Electrical Engineering and Technology
• /
• v.11 no.2
• /
• pp.265-273
• /
• 2016

A water wall system is one of the most important components of a boiler in a thermal power plant, and it is a nonlinear Multi-Input and Multi-Output (MIMO) system, with 6 inputs and 3 outputs. Three models are developed and comp for the controller design, including a linear model, a multilayer feed-forward neural network (MFNN) model and an Echo State Network (ESN) model. First, the linear model is developed by linearizing a given nonlinear model and is analyzed as a function of the operating point. Second, the MFNN and the ESN are developed by using training data from the nonlinear model. The three models are validated using Matlab with nonlinear input-output data that was not used during training.