• Earthquakes and Structures
• /
• v.24 no.1
• /
• pp.1-9
• /
• 2023

A model for calculating structure interacted mechanics is proposed. A structural interaction model and controller design based on tuned mass damping (TMD) was developed to control the induced vibration. A key point is to introduce a new analytical model to evaluate the properties of the TMD that recognizes the motion-dependent nonlinear response observed in the simulations. Aiming at the problem of increased current harmonics and low efficiency of permanent magnet synchronous motors for electric vehicles due to dead time effect, a dead time compensation method based on neural network filter and current polarity detection is proposed. Firstly, the DC components and the higher harmonic components of the motor currents are obtained by virtue of what the neural network filters and the extracted harmonic currents are adjusted to the required compensation voltages by virtue of what the neural network filters. Then, the extracted DC components are used for current polarity dead time compensation control to avert the false compensation when currents approach zero. The neural network filter method extracts the required compensation voltages from the speed component and the current polarity detection compensation method obtains the required compensation voltages by discriminating the current polarity. The combination of the two methods can more precisely compensate the dead time effect of the control system to improve the control performance. Furthermore, based on the relaxed method, the intelligent approach of stability criterion can be regulated appropriately and the artificial TMD was found to be effective in reducing cross-wind vibrations.

• Proceedings of the KOR-KST Conference
• /
• 2003.02a
• /
• pp.127-150
• /
• 2003

This research presents a novel application of static traffic assignment methods, but with a variable time value, for estimating the market share of a high-speed rail (HSR) in the NW-SE corridor of Korea which is currently served by the airline (AR), conventional rail (CR), and highway (HWY) modes. The proposed model employs the time-space network structure to capture the interrelations among all competing transportation modes, and to reflect their supply- and demand-sides constraints as well as interactions through properly formulated link-node structures. The embedded cost function for each network link offers the flexibility for incorporating all associated factors, such as travel time and fare, in the model computation, and enables the use of a distribution rather than a constant to represent the time-value variation among all transportation mode users. To realistically capture the tripmakers' value-of-time (VOT) along the target area, a novel method for VOT calibration has been developed with aggregate demand information and key system performance data from the target area. Under the assumption that intercity tripmakers often have nearly "perfect" travel information, one can solve the market share of each mode after operations of HSR for each O-D pair under the time-dependent demand with state-of-the-art traffic assignment. Aside from estimating new market share, this paper also investigated the impacts of HSR on other existing transportation modes.

• Speech Sciences
• /
• v.3
• /
• pp.119-131
• /
• 1998

We have developed a keyword spotting system for automatic speech editing. This system recognizes the only keyword 'MBC news' and then sends the time information to the host system. We adopted a vocabulary dependent model based on continuous hidden Markov model, and the Viterbi search was used for recognizing the keyword. In recognizing the keyword, the system uses a parallel network where HMM models are connected independently and back-tracking information for reducing false alarms and missing. We especially focused on implementing a stable and practical real-time system.

• Journal of Korea Multimedia Society
• /
• v.6 no.7
• /
• pp.1209-1218
• /
• 2003

To guarantee client buffer stabilization and visual quality, the VOD service that provides real time video titles on requirements of numerous users, needs a mechanism which transfers frames with dropping or skipping algorithm by network condition. In this paper, we show an algorithm that transfers withdrawed skipped MPEG frames(I, P, B frame) from disk to client dependent on network load. Moreover, we verify through a simulation that adaptive dealing on network load can reduce the network load and stabilize client receiving buffer.

• 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.

• Journal of Intelligence and Information Systems
• /
• v.2 no.2
• /
• pp.69-83
• /
• 1996

The objective of this research is to develop a knowledge acquisition and refinement method for a multi-objective and multi-decision FMS scheduling problem. A competitive neural network and an inductive learning algorithm are integrated to extract and refine necessary scheduling knowledge from simulation outputs. The obtained scheduling knowledge can assist the FMS operator in real-time to decide multiple decisions simultaneously, while maximally meeting multiple objective desired by the FMS operator. The acquired scheduling knowledge for an FMS scheduling problem is tested by comparing the desired and the simulated values of the multiple objectives. The result show that the knowledge acquisition and refinement method is effective for the multi-objective and multi-decision FMS scheduling problems.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.59 no.3
• /
• pp.97-110
• /
• 2017

This study investigates the efficiencies of machine learning models, including artificial neural network (ANN), generalized regression neural network (GRNN), adaptive neuro-fuzzy inference system (ANFIS) and random forest (RF), for reservoir water level forecasting in the Chungju Dam, South Korea. The models' efficiencies are assessed based on model efficiency indices and graphical comparison. The forecasting results of the models are dependent on lead times and the combination of input variables. For lead time t = 1 day, ANFIS1 and ANN6 models yield superior forecasting results to RF6 and GRNN6 models. For lead time t = 5 days, ANN1 and RF6 models produce better forecasting results than ANFIS1 and GRNN3 models. For lead time t = 10 days, ANN3 and RF1 models perform better than ANFIS3 and GRNN3 models. It is found that ANN model yields the best performance for all lead times, in terms of model efficiency and graphical comparison. These results indicate that the optimal combination of input variables and forecasting models depending on lead times should be applied in reservoir water level forecasting, instead of the single combination of input variables and forecasting models for all lead times.

• KNOM Review
• /
• v.22 no.2
• /
• pp.1-11
• /
• 2019

The paradigm of science and technology and R&D system changes from the existing form of independent research to community research which transmits and stores large-capacity data generated from observation and experiment to the research site through a high-speed network. Network-dependent advanced research is improves form of collaboration, sharing, globalization. For this purpose, not only the integration of science and IT technology, but also the optimized network infrastructure and technology for real-time sharing of the enormous data is necessary. This paper explains the direction of change through the case of KREONET's network-dependent advanced research and analyzes the technology and network resources supported in each research field. Through this, we propose a network optimization model for each advanced research field with the main data flow and satisfy the network required performance by advanced applications on KREONET.

• Journal of Environmental Science International
• /
• v.15 no.2
• /
• pp.141-155
• /
• 2006

This study is aimed at the development of a runoff forecasting model to solve the uncertainties occurring in the process of rainfall-runoff modeling and improve the modeling accuracy of the stream runoff forecasting, The study area is the downstream of Naeseung-chun. Therefore, time-dependent data was obtained from the Wolpo water level gauging station. 11 and 2 out of total 13 flood events were selected for the training and testing set of model. The model performance was improved as the measuring time interval$(T_m)$ was smaller than the sampling time interval$(T_s)$. The Neuro-Fuzzy(NF) and TANK models can give more accurate runoff forecasts up to 4 hours ahead than the Feed Forward Multilayer Neural Network(FFNN) model in standard above the Determination coefficient$(R^2)$ 0.7.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2002.10a
• /
• pp.1062-1065
• /
• 2002

We have evaluated the ability of a time-delayed artificial neural network (TDANN) to predict muscle forces using only eletromyographic(EMG) signals. To achieve this goal, tendon forces and EMG signals were measured simultaneously in the gastrocnemius muscle of a dog while walking on a motor-driven treadmill. Direct measurements of tendon forces were performed using an implantable force transducer and EMG signals were recorded using surface electrodes. Under dynamic conditions, the relationship between muscle force and EMG signal is nonlinear and time-dependent. Thus, we adopted EMG amplitude estimation with adaptive smoothing window length. This approach improved the prediction ability of muscle force in the TDANN training. The experimental results indicated that dynamic tendon forces from EMG signals could be predicted using the TDANN, in vivo.