• The KIPS Transactions:PartB
• /
• v.13B no.2 s.105
• /
• pp.133-138
• /
• 2006

Learning with Attribute Value Taxonomies (AVT) has shown that it is possible to construct accurate, compact and robust classifiers from a partially missing dataset (dataset that contains attribute values specified with different level of precision). Yet, in many cases AVTs are generated from experts or people with specialized knowledge in their domain. Unfortunately these user-provided AVTs can be time-consuming to construct and misguided during the AVT building process. Moreover experts are occasionally unavailable to provide an AVT for a particular domain. Against these backgrounds, this paper introduces an AVT generating method called GA-AVT-Learner, which finds a near optimal AVT with a given training dataset using a genetic algorithm. This paper conducted experiments generating AVTs through GA-AVT-Learner with a variety of real world datasets. We compared these AVTs with other types of AVTs such as HAC-AVTs and user-provided AVTs. Through the experiments we have proved that GA-AVT-Learner provides AVTs that yield more accurate and compact classifiers and improve performance in learning missing data.

• The Journal of Korea Robotics Society
• /
• v.17 no.3
• /
• pp.365-372
• /
• 2022

The accurate object recognition is important for the precise and accurate manipulation. To enhance the recognition performance, we can use various types of sensors. In general, acquired data from sensors have a high sampling rate. So, in the past, the RNN-based model is commonly used to handle and analyze the time-series sensor data. However, the RNN-based model has limitations of excessive parameters. CNN-based model also can be used to analyze time-series input data. However, CNN-based model also has limitations of the small receptive field in early layers. For this reason, when we use a CNN-based model, model architecture should be deeper and heavier to extract useful global features. Thus, traditional methods like RN N -based and CN N -based model needs huge amount of learning parameters. Recently studied result shows that Fast Fourier Convolution (FFC) can overcome the limitations of traditional methods. This operator can extract global features from the first hidden layer, so it can be effectively used for feature extracting of sensor data that have a high sampling rate. In this paper, we propose the algorithm to recognize objects using tactile sensor data and the FFC model. The data was acquired from 11 types of objects to verify our posed model. We collected pressure, current, position data when the gripper grasps the objects by random force. As a result, the accuracy is enhanced from 84.66% to 91.43% when we use the proposed FFC-based model instead of the traditional model.

• Proceedings of the KIEE Conference
• /
• 1996.07a
• /
• pp.139-141
• /
• 1996

The switched reluctance motors(SRM) are simple and robust in structure. Because the wide range of power and speed, their application field is increasing. In order to design the motors and to evaluate the performance of them properly, an accurate study about the analysis of motor characteristics is required. In this paper, for the analysis of SRM characteristics, the finite element method which is based on the solution of combined equations both the electromagnetic field equations and the circuit equations of stator is adopted. The analysis model is to he assumed two-dimensional and the nonlinear property of magnetic materials is considered by Newton-Raphson method. To verify the usefulness of the proposed algorithm, commercial SRM is chosen and simulated. The computed torques obtained by Maxwell Stress Tensor are compared with the experimental data and it is found that they are in good agreement. By applying the proposed algorithm to two cases, currents of stator and torques at every angular positions of rotor are obtained step by step. Comparing them, one can recognize that torque ripple of SRM can he improved by controlling the switching sequences of driving circuits.

• Journal of Institute of Control, Robotics and Systems
• /
• v.20 no.4
• /
• pp.395-401
• /
• 2014

Magnetic levitation systems using the attraction force of electromagnets have many constraints according to the variation of air gap and the nonlinearity of electromagnetic force and inductances. As a result of these constraints, the nonlinear control of a magnetic levitation system has been improved by the latest advanced processors and accurate measurement system which can overcome problems such as many constraints and nonlinearity. This paper concentrates on the modeling of a nonlinear magnetic levitation system and an application of an exponential reaching law based sliding mode controller using the exponential reaching law which is one of the most robust controllers against external unexpected disturbances or parameter fluctuations. Controllability of a magnetic levitation system using the sliding mode control algorithm and robustness against parameter fluctuations have been verified through the experimental results.

• Journal of electromagnetic engineering and science
• /
• v.1 no.1
• /
• pp.105-111
• /
• 2001

This paper presents a new and robust technique for a coarse frequency offset estimation in OFDM systems. As an evaluation of the proposed algorithm, we apply it to Eureka 147 DAB system. The proposed coarse frequency offset estimation algorithm is based on the differential detection technique between adjacent subcarriers to eliminate the phase shift effects of symbol timing offset and fractional frequency offset. A coarse frequency offset is determined from the correlation output between a received interarrier differential phase reference symbol and several locally generated but frequency-shifted intercarrier differential phase reference symbols. The performance of our estimation algorithm is evaluated by means of computer simulation and is compared with those of previous proposed algorithms for DAB transmission modes I, II, III, and IV. Simulation results show that the proposed algorithm generates extremely accurate estimates with low complexity irrespective of the symbol timing offset.

• Nuclear Engineering and Technology
• /
• v.55 no.9
• /
• pp.3301-3312
• /
• 2023

This study incorporates a high-fidelity transient analysis solver based on multigroup CMFD in the MOC code STREAM. Transport modeling with heterogeneous geometries of the reactor core increases computational cost in terms of memory and time, whereas the multigroup CMFD reduces the computational cost. The reactor condition does not change at every time step, which is a vital point for the utilization of CMFD. CMFD correction factors are updated from the transport solution whenever the reactor core condition changes, and the simulation continues until the end. The transport solution is adjusted once CMFD achieves the solution. The flux-weighted method is used for rod decusping to update the partially inserted control rod cell material, which maintains the solution's stability. A smaller time-step size is needed to obtain an accurate solution, which increases the computational cost. The adaptive step-size control algorithm is robust for controlling the time step size. This algorithm is based on local errors and has the potential capability to accept or reject the solution. Several numerical problems are selected to analyze the performance and numerical accuracy of parallel computing, rod decusping, and adaptive time step control. Lastly, a typical pressurized LWR was chosen to study the rod-ejection accident.

• Journal of the Korean Society of Safety
• /
• v.39 no.2
• /
• pp.28-37
• /
• 2024

The recent surge in the production and handling of hazardous materials in Korea necessitates developing and implementing robust emergency response plans. These plans are crucial in safeguarding the well-being of workers and residents in the event of an incident. The consequence analysis methodology outlined in the KOSHA guidelines provides a foundation for designing emergency response plans in the event of chemical accidents. However, the consequence analysis is evaluated based on assumed accident cases or worst-case scenarios. Consequently, the emergency response plan based on the consequence analysis may overestimate the damage area, complicating rescue efforts and unnecessarily increasing costs. More information and parameters become available after an accident, enabling more accurate consequence analysis. This implies that the results of consequence analysis based on this detailed information provide more realistic results than those based on assumed accidents. This study attempts to optimize the resource allocation and cost-effectiveness of emergency response plans for chemical accidents. Existing procedures and manuals are revised to elucidate the proposed model and conduct real-time consequence analysis. The existing emergency response plan is compared to verify the proposed model's efficacy. The obtained results indicate that the proposed model can exhibit better performance.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.18 no.10
• /
• pp.997-1005
• /
• 2008

Linear motor stage is a useful device in precision engineering field because of its simple power transmission mechanism and accurate positioning. Even though linear motor stage shows fine positioning accuracy along travel axis, geometric dependent errors which relay on machining and assembling accuracy should be addressed to increase total positioning performances. In this paper, we suggests a cost effective yaw error compensation servo-system which is mounted on platform of the stage and nullify travel position dependent yaw error. This paper also provides a method of designing a sliding mode control which is robust to existing friction disturbance and model uncertainties. The reachability condition of slinding mode control for the yaw error compensating servo-system has been established. From some experimental results by using an experimental set-up, the sliding mode control showed its effective in disturbance rejection and its performance was superior to conventional linear controls.

• Journal of Internet Computing and Services
• /
• v.13 no.3
• /
• pp.107-114
• /
• 2012

In this paper, we proposed an automatic monitoring system for contact center in order to manage customer's complaint and agent's quality. The proposed system allows more accurate monitoring using emotion recognition and keyword spotting method for neutral/anger voice emotion. The system can provide professional consultation and management for the customer with language violence, such as abuse and sexual harassment. We developed a method of building robust algorithm on heterogeneous speech DB of many unspecified customers. Experimental results confirm the stable and improved performance using real contact center speech data.

• Journal of Korea Multimedia Society
• /
• v.21 no.8
• /
• pp.876-887
• /
• 2018

In this paper, we propose an robust CAMShift method to track stage objects with a camera. In order to solve the problem of tracking object misdetection in existing CAMShift technique, MBR region is detected to separate the background and the subject, and the subject size of the region of interest is calculated to solve the problem of erroneously detecting a large region having a similar color distribution ratio. Also, by applying the color corelogram and MB-LBP to the part that can not be solved by the color ratio and the size limitation, accurate texture tracking is enabled by reflecting the texture characteristics. Experimental results show that the proposed method has good tracking performance for objects that do not deviate from the size of the subject set in the area of interest and accurately extracts the texture characteristics of different subjects with similar color distribution ratios.