• 제어로봇시스템학회:학술대회논문집
• /
• 1997.10a
• /
• pp.422-425
• /
• 1997

We consider the robust control problem for nonminimum phase(NMP) systems with parametric uncertainty which appear often in aircraft and missile control. First, a new method that makes such an uncertain NMP system to be factored as a interval minimum phase(MP) transfer function and a time delay term in the Pade approximation form has been presented. The controller to be proposed consists of a compensator $C_{Q}$(s) with Smith predictor in the internal model control(IMC) structure, so that it can have good robustness and performance against the structured uncertainty and the time delay behaviour due to NMP plant the $C_{Q}$(s) is designed on the MP model by using QFT. The stability and performance of overall system has been evaluated by the generalized Kharitonov theorem.rem.

• Journal of the Korean Mathematical Society
• /
• v.48 no.4
• /
• pp.775-795
• /
• 2011

The disturbance attenuation problem for a class of discretetime switched linear systems with exponential uncertainties via switched state feedback and switched dynamic output feedback is investigated, respectively. By using Taylor series approximation and convex polytope technique, exponentially uncertain discrete-time switched linear system is transformed into an equivalent switched polytopic model with additive norm bounded uncertainty. For such equivalent switched model, one designs its switching strategy and associated state feedback controllers and dynamic output feedback controllers so that whole switched model is asymptotical stabilization with H-in nity disturbance attenuation base on switched Lyapunov function and LMI approach. Finally, two numerical examples are presented to illustrate our results.

• Smart Structures and Systems
• /
• v.19 no.4
• /
• pp.431-439
• /
• 2017

This investigation is aimed to develop a model of experimental-computation determination of a support moment of a cantilever beam loaded with concentrated force at its end including the optimal choice of coordinates of deflection data points and parameters of transformation of deflection data in case of insufficient accuracy of the assignment of initial parameters (support settlement, angle of rotation of the bearing section) and cantilever beam length. The influence of distribution and characteristics of sensors on the cantilever beam on the accuracy of determining the support moment which improves in the course of transition from the uniform distribution of sensors to optimal non-uniform distribution is shown. On the basis of the theory of inverse problems the method of transformation reduction at numerical differentiation of deflection functions has been studied. For engineering evaluation formulae of uncertainty estimate to determine a support moment of a cantilever beam at predetermined uncertainty of measurements using sensors have been obtained.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.30B no.4
• /
• pp.92-103
• /
• 1993

In this paper, we propose an adaptive classifier based on uncertainty of features for 3D planar object recognition. First, we investigate the uncertainty of depth information and the feature values of 3D planar object by numerical method. And, we observed that the statistical behavior of feature is dependent on the position and orientation of objects. After that, the approximation of the statistical behavior is executed. Subsequently, the recognition procedure is executed by the adaptive classifier. By computer simulation, we confirmed that the proposed classifier is useful for 3D planar object recognition.

• Proceedings of the Korean Operations and Management Science Society Conference
• /
• 1996.04a
• /
• pp.525-529
• /
• 1996

Decision environments involve a high degree of uncertainty as well as multiple, conflicting goals. Although traditional goal programming offers a means of considering multiple, conflicting goals and arrives at a satisficing solution in a deterministic manner, its major drawback is that decision makers often specify aspiration level of each goal as a single number. To overcome the problem of setting aspiration levels, chance constrained programming can be incorporated into goal programming formulation so that sampling information can be utilized to describe uncertainty distribution. Another drawback of goal programming is that it does not provide a systematic approach to set priorities and trade-offs among conflicting goals. To overcome this weekness, the analytic hierarchy process(AHP) is used in the model. Also, most goal programming models in the literature are of a linear form, although some nonlinear models have been presented. Consideration of risk in technological coefficients and right hand sides, however, leads to nonlinear goal programming models, which require a linear approximation to be solved. In this paper, chance constrained reformulation with linear approximation is presented for a 0-1 goal programming problem whose technological coefficients and right hand sides are stochastic. The model is presented with a numerical example for the purpose of demonstration.

• Journal of KIISE:Software and Applications
• /
• v.31 no.11
• /
• pp.1431-1438
• /
• 2004

Reinforcement learning learns policies for accomplishing a task's goal by experience through interaction between agent and environment. Q-learning, basis algorithm of reinforcement learning, has the problem of curse of dimensionality and slow learning speed in the incipient stage of learning. In order to solve the problems of Q-learning, new function approximation methods suitable for reinforcement learning should be studied. In this paper, to improve these problems, we suggest Fuzzy Q-Map algorithm that is based on online fuzzy clustering. Fuzzy Q-Map is a function approximation method suitable to reinforcement learning that can do on-line teaming and express uncertainty of environment. We made an experiment on the mountain car problem with fuzzy Q-Map, and its results show that learning speed is accelerated in the incipient stage of learning.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.31 no.4
• /
• pp.490-495
• /
• 2007

This research aims to examine accuracy and efficiency of the first four moments corresponding to mean, standard deviation, skewness, and kurtosis, which are estimated by a function approximation moment method (FAMM). In FAMM, the moments are estimated from an approximating quadratic function of a system response function. The function approximation is performed on a specially selected experimental region for accuracy, and the number of function evaluations is taken equal to that of the unknown coefficients for efficiency. For this purpose, three error-minimizing conditions are utilized and corresponding canonical experimental regions constructed accordingly. An interpolation function is then obtained using a D-optimal design and then the first four moments of it are obtained as the estimates for the system response function. In order to verify accuracy and efficiency of FAMM, several non-linear examples are considered including a polynomial of order 4, an exponential function, and a rational function. The moments calculated from various coefficients of variation show very good accuracy and efficiency in comparison with those from analytic integration or the Monte Carlo simulation and the experimental design technique proposed by Taguchi and updated by D'Errico and Zaino.

• Journal of the Korean Chemical Society
• /
• v.47 no.1
• /
• pp.31-37
• /
• 2003

The expanded uncertainties calculated by the application of GUM -approximation and Monte-Carlo simulation were compared about the model equation of one-point calibration which is widely used for the measurements and chemical analysis. For the comparisons, we assumed a set of artificial data at the various level of concentration and dispersion of t or normal distribution. Mistakes of more then 50 % was revealed at the values calculated by GUM-approximation in comparison with those of Monte-Carlo simulation because of the excess dispersion from t-distribution and non-linearity by division in the equation. In contrary, the mistake of calculation due to non-linearity of the equation was not observed in the level of detection limits with the equation of one-point calibration, because of the relatively large values of uncertainty in response.

• Nuclear Engineering and Technology
• /
• v.50 no.7
• /
• pp.1043-1050
• /
• 2018

For an efficient Monte Carlo (MC) burnup analysis, an accurate high-order depletion scheme to consider the nonlinear flux variation in a coarse burnup-step interval is crucial accompanied with an accurate depletion equation solver. In a Seoul National University MC code, McCARD, the high-order depletion schemes of the quadratic depletion method (QDM) and the linear extrapolation/quadratic interpolation (LEQI) method and a depletion equation solver by the Chebyshev rational approximation method (CRAM) have been newly implemented in addition to the existing constant extrapolation/backward extrapolation (CEBE) method using the matrix exponential method (MEM) solver with substeps. In this paper, the quadratic extrapolation/quadratic interpolation (QEQI) method is proposed as a new high-order depletion scheme. In order to examine the effectiveness of the newly-implemented depletion modules in McCARD, four problems in the VERA depletion benchmarks are solved by CEBE/MEM, CEBE/CRAM, LEQI/MEM, QEQI/MEM, and QDM for gadolinium isotopes. From the comparisons, it is shown that the QEQI/MEM predicts ${k_{inf}}^{\prime}s$ most accurately among the test cases. In addition, statistical uncertainty propagation analyses for a VERA pin cell problem are conducted by the sensitivity and uncertainty and the stochastic sampling methods.

• Journal of Korea Water Resources Association
• /
• v.43 no.4
• /
• pp.337-351
• /
• 2010

There is a growing dissatisfaction with use of conventional statistical methods for the prediction of extreme events. Conventional methodology for modeling extreme event consists of adopting an asymptotic model to describe stochastic variation. However asymptotically motivated models remain the centerpiece of our modeling strategy, since without such an asymptotic basis, models have no rational for extrapolation beyond the level of observed data. Also, this asymptotic models ignored or overestimate the uncertainty and finally decrease the reliability of uncertainty. Therefore this article provide the research example of the extreme rainfall event and the methodology to reduce the uncertainty. In this study, the Bayesian MCMC (Bayesian Markov Chain Monte Carlo) and the MLE (Maximum Likelihood Estimation) methods using a quadratic approximation are applied to perform the at-site rainfall frequency analysis. Especially, the GEV distribution and Gumbel distribution which frequently used distribution in the fields of rainfall frequency distribution are used and compared. Also, the results of two distribution are analyzed and compared in the aspect of uncertainty.