• 제어로봇시스템학회:학술대회논문집
• /
• 1996.10b
• /
• pp.192-194
• /
• 1996

This paper propose an optimal scaling gain tuning method of the fuzzy PI controller using Genetic Algorithm(GA). Scaling gains can reflect the control resolution and fuzziness of input/output variables. By the scaling gain method, the design of a fuzzy logic controller(FLC) can be simplified without affecting the system performance in comparison with multi-decision table method. In designing a fuzzy logic controller, the analytic approach method for the optimization is unavailable. Therefore GA is excellent optimization algorithms for scaling gain tuning. Using this optimal scaling gain tuning method, a good performance can be achieved both in transient and steady state.

• Journal of the Korean Society for Precision Engineering
• /
• v.26 no.12
• /
• pp.55-61
• /
• 2009

Machining systems have been evolved to produce more detailed products of high added value. This has been possible, in large part, due to the development of highly accurate multi-axis CNC machine tools. The conventional CNC of machine tools has individual axis controllers to maximize tracking performance. On the other hand, cross-coupling controllers can be integrated into the conventional CNC to enhance contouring performance. For this multi-axis cross-coupling control system, it is necessary to automatically adjust the controller gains depending on operating conditions and/or other external conditions from an optimization perspective. This paper proposes automatic modeling of feed drive systems that minimizes the difference in behavior between the system model and the actual system. Based on the modeling, an integrated auto-tuning method is also proposed to improve both tracking and contouring accuracy of a 3-axis cross-coupling control system as well as users' convenience. The proposed methods are evaluated by both simulation and experiments.

• Journal of Internet Computing and Services
• /
• v.10 no.4
• /
• pp.199-211
• /
• 2009

Recent advancements in wireless networks have enabled support for mobile applications to transfer data over heterogeneous wireless paths in parallel using data striping technique [2]. Traditionally, high performance data transfer requires tuning of multiple TCP sockets, at sender's end, based on bandwidth delay product (BDP). Moreover, traditional techniques like Automatic TCP Buffer Tuning (ATBT), which balance memory and fulfill network demand, is designed for wired infrastructure assuming single flow on a single socket. Hence, in this paper we propose a buffer tuning technique at senders end designed to ensure high performance data transfer by striping data at transport layer across heterogeneous wireless paths. Our mechanism has the capability to become a resource management system for transport layer connections running on multi-homed mobile host supporting features for wireless link i.e. mobility, bandwidth fluctuations, link level losses. We show that our proposed mechanism performs better than ATBT, in efficiently utilizing memory and achieving aggregate throughput.

• Journal of the Korea Society of Computer and Information
• /
• v.24 no.11
• /
• pp.51-59
• /
• 2019

Due to the exponential growth of access information on the web, the need for predicting web users' next access has increased. Various models such as markov models, deep neural networks, support vector machines, and fuzzy inference models were proposed to handle web access prediction. For deep learning based on neural network models, training time on large-scale web usage data is very huge. To address this problem, deep neural network models are trained on cluster of computers in parallel. In this paper, we investigated impact of several important spark parameters related to data partitions, shuffling, compression, and locality (basic spark parameters) for training Multi-Layer Perceptron model on Spark standalone cluster. Then based on the investigation, we tuned basic spark parameters for training Multi-Layer Perceptron model and used it for tuning Spark when training Multi-Layer Perceptron model for web access prediction. Through experiments, we showed the accuracy of web access prediction based on our proposed web access prediction model. In addition, we also showed performance improvement in training time based on our spark basic parameters tuning for training Multi-Layer Perceptron model over default spark parameters configuration.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.43 no.2 s.344
• /
• pp.34-41
• /
• 2006

This paper presents a CMOS multi-mode baseband filter architecture to support PDC/GSM/EDGE/WCDMA and its new automatic tuning method. 5-th order Chebyshev low pass filter is designed for implementing the baseband channel-select filter. Capacitors and resistors were shared efficiently between modes to minimize the area. And, the new cut-off frequency tuning method is proposed to compensate the process variation. This method can reduce the area and the noise level due to MOS switches.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.44 no.6
• /
• pp.21-28
• /
• 2007

Using multi-resolution, the mutiresolution proportional-integral-derivative(MRPID) controller functions as a filter to eliminate noise and disturbance which are included in error signals. If the sampling frequency is high, the response time will be delayed because of the remaining high frequency component although the overshoot is removed. However, if the sampling frequency is low, the response time will be enhanced by getting rid of signal components while the overshoot is increased. In this paper, the sampling frequency tuning method is used the response of the proportional integral derivative(PID) controller and the MRPID controller, and the parameter tuning method is considered the characteristic of the MRPID controller. The proposal method is verified by computer simulations.

• Journal of KSNVE
• /
• v.11 no.1
• /
• pp.147-155
• /
• 2001

This paper is concerned with the real-time automatic tuning of the multi-input multi-output positive position feedback controllers for smart structures by the genetic algorithms. The genetic algorithms have proven its effectiveness in searching optimal design parameters without falling into local minimums thus rendering globally optimal solutions. The previous real-time algorithm that tunes a single control parameter is extended to tune more parameters of the MIMO PPF controller. We employ the MIMO PPF controller since it can enhance the damping value of a target mode without affecting other modes if tuned properly. Hence, the traditional positive position feedback controller can be used in adaptive fashion in real time. The final form of the MIMO PPF controller results in the centralized control, thus it involves many parameters. The bounds of the control Parameters are estimated from the theoretical model to guarantee the stability. As in the previous research, the digital MIMO PPF control law is downloaded to the DSP chip and a main program, which runs genetic algorithms in real time, updates the parameters of the controller in real time. The experimental frequency response results show that the MIMO PPF controller tuned by GA gives better performance than the theoretically designed PPF. The time response also shows that the GA tuned MIMO PPF controller can suppress vibrations very well.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.05a
• /
• pp.182-185
• /
• 2003

This paper presents a new design algorithm for the combination with the fuzzy classifier and the Bayesian classifier. Only few attempts have so far been made at providing an effective design algorithm combining the advantages and removing the disadvantages of two classifiers. Specifically, the suggested algorithms are composed of three steps: the combining, the fuzzy-set-based pruning, and the fuzzy set tuning. In the combining, the multi-inputs and multi-outputs (MIMO) fuzzy model is used to combine two classifiers. In the fuzzy-set-based pruning, to effectively decrease the complexity of the fuzzy-Bayesian classifier and the risk of the overfitting, the analysis method of the fuzzy set and the recursive pruning method are proposesd. In the fuzzy set tuning for the misclassified feature vectors, the premise parameters are adjusted by using the gradient decent algorithm. Finally, to show the feasibility and the validity of the proposed algorithm, a computer simulation is provided.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.11
• /
• pp.39-46
• /
• 2022

In this paper, we propose a multi-task model that can simultaneously predict general-purpose tasks such as part-of-speech tagging, lemmatization, and dependency parsing using the UD Korean Kaist v2.3 corpus. The proposed model thus applies the self-attention technique of the BERT model and the graph-based Biaffine attention technique by fine-tuning the multilingual BERT and the two Korean-specific BERTs such as KR-BERT and KoBERT. The performances of the proposed model are compared and analyzed using the multilingual version of BERT and the two Korean-specific BERT language models.

• Journal of the Korean Society of Industry Convergence
• /
• v.3 no.2
• /
• pp.115-121
• /
• 2000

This paper presents reducing method of profile error on a mechanical tuning for multi-spindle control of NC machine tools. To reduce the profile error in the feed drive system, it is useful to adopt same transfer function of multi spindle machine tools. By selecting the correction vector of servo rigidity and natural vibration on JK map, multi spindle control can be tuned by mechanical parameters with small profile error.