• The Journal of the Acoustical Society of Korea
• /
• v.22 no.1E
• /
• pp.19-25
• /
• 2003

Ryu et al. recently proposed a multiple target angle-tracking algorithm without a data association problem. This algorithm, however, shows the degraded performance on evasive maneuvering targets, because the estimated signal subspace is d,:graded in the algorithm. In this Paper, we proposed a new algorithm, in which VFF-PASTd (Variable Forgetting Factor PASTd) algorithm is applied to Ryu's algorithm to effectively handle the evasive target tracking with better time-varying signal subspace.

• Proceedings of the IEEK Conference
• /
• 2003.07d
• /
• pp.1251-1254
• /
• 2003

In this paper, Experience Sensitive Cumulative Neural Network (ESCNN) is introduced, which can cumulate the same or similar experiences. As the same or similar training patterns are cumulated in the network, the system recognize more important information in the training patterns. The functions of forgetting less important informations and attending more important informations resided in the training patterns are surveyed and implemented by simulations. The system behaves well under the noisy circumstances due to its forgetting and/or attending properties, even in 50 percents noisy environments. This paper also describes the creation of the generalized patterns for the input training patterns.

• 제어로봇시스템학회:학술대회논문집
• /
• 1987.10a
• /
• pp.817-826
• /
• 1987

An offset-free self-tuning control with pole placement (STCPP) and a recursive parameter estimation with multiple and variable forgetting factors (REWF), together with their application to a real plant, are described. There are two different types of offset-free STCPP; their features are analysed and discussed. REMVF employs as many forgetting factors as parameter estimates. It is suitable when parameters to be estimated are changing at different rates. The offset-free STCPP and REMVF have been successfully applied to a real plant, giving excellent results.

• Journal of Drive and Control
• /
• v.13 no.4
• /
• pp.59-67
• /
• 2016

This paper proposes a model-prediction-based collision-avoidance algorithm for excavators for which the recursive-least-squares (RLS) estimation of the excavator's rotational inertia is used. To estimate the rotational inertia of the excavator, the RLS estimation with multiple forgetting and two updating rules for the nominal parameter and the forgetting factors was conducted based on the excavator-swing dynamics. The average value of the estimated rotational inertia that is for the minimizing effects of the estimation error was computed using the recursive-average method with forgetting. Based on the swing dynamics, the computed average of the rotational inertia, the damping coefficient for braking, and the excavator's braking angle were predicted, and the predicted braking angle was compared with the detected-object angle for a safety evaluation. The safety level defined in this study consists of the three levels safe, warning, and emergency braking. The analytical rotational-inertia-based performance evaluation of the designed estimation algorithm was conducted using a typical working scenario. The results of the safety evaluation show that the predictive safety-evaluation algorithm of the proposed model can evaluate the safety level of the excavator during its operation.

• Korean Journal of Psychosomatic Medicine
• /
• v.5 no.1
• /
• pp.31-42
• /
• 1997

This study was aimed to investigate dysfunctional attitudes, stress coping strategies and depressive symptoms in psychiatric patients. The subjects of this study consisted of 210 patients(138 schizophrenic patients, 29 depression patients, 43 alcohol dependence patients) according to DSM-IV criteria. Futhermore, the instruments were K-BDI(Beck Depression Inventory-Korean version), DAS(Dysfunctional Attitude Scale) and multidimensional coping strategy scale. The results were the following. 1) There were statistically significant correlations between depressive symptoms and dysfunctional attitudes in psychiatric patients. 2) In terms of coping strategies, there were positive correlations between depressive symptoms and focus on and venting emotions, accommodation, active forgetting, self-criticism, positive comparison, fatalism, passive withdrawal. Whereas, there was significant negative correlation between depressive symptom and active coping. 3) In terms of coping strategies, there were significant correlations between dysfunctional attitudes and focus on and venting emotions, active forgetting, self-criticism, positive comparison, fatalism, passive withdrawal. 4) Depression groups reported significantly higher BDI scores than schizophrenia groups. 5) In depression groups, DAS scores were significantly higher than those in schizophrenia groups. 6) In terms of coping strategies according to diagnosis, there were significant differences in venting emotions, active forgetting and self-criticism. As for venting emotions, alcoholic groups were scored significantly higher than schizophrenic groups. As for active forgetting, depression groups were scored significantly higher than schizophrenic groups. In self-criticism, depression groups and alcohol dependence groups reported significantly higher scores than schizophrenic groups.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.1
• /
• pp.27-35
• /
• 2021

This paper proposes a method to mitigate the Catastrophic Forgetting phenomenon in which artificial neural networks forget information on previous data. This method adjusts the Regularization strength by measuring the relationship between previous data and present data. MNIST and EMNIST data were used for performance evaluation and experimented in three scenarios. The experiment results showed a 0.1~3% improvement in the accuracy of the previous task for the same domain data and a 10~13% improvement in the accuracy of the previous task for different domain data. When continuously learning data with various domains, the accuracy of all previous tasks achieved more than 50% and the average accuracy improved by about 7%. This result shows that neural network learning can be properly performed in a CL environment in which data of different domains are successively entered by the method of this paper.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.1
• /
• pp.15-23
• /
• 2022

Among the learning methods for Continuous Learning environments, "Learning without Forgetting" has fixed regularization strengths, which can lead to poor performance in environments where various data are received. We suggest a way to set weights variable by identifying the features of the data we want to learn. We applied weights adaptively using correlation and complexity. Scenarios with various data are used for evaluation and experiments showed accuracy increases by up to 5% in the new task and up to 11% in the previous task. In addition, it was found that the adaptive weight value obtained by the algorithm proposed in this paper, approached the optimal weight value calculated manually by repeated experiments for each experimental scenario. The correlation coefficient value is 0.739, and overall average task accuracy increased. It can be seen that the method of this paper sets an appropriate lambda value every time a new task is learned, and derives the optimal result value in various scenarios.

• Smart Structures and Systems
• /
• v.32 no.1
• /
• pp.9-21
• /
• 2023

The parameters of civil engineering structures have time-variant characteristics during their service. When extremely large external excitations, such as earthquake excitation to buildings or overweight vehicles to bridges, apply to structures, sudden or gradual damage may be caused. It is crucially necessary to detect the occurrence time and severity of the damage. The unscented Kalman filter (UKF), as one efficient estimator, is usually used to conduct the recursive identification of parameters. However, the conventional UKF algorithm has a weak tracking ability for time-variant structural parameters. To improve the identification ability of time-variant parameters, an adaptive UKF with forgetting factor (AUKF-FF) algorithm, in which the state covariance, innovation covariance and cross covariance are updated simultaneously with the help of the forgetting factor, is proposed. To verify the effectiveness of the method, this paper conducted two case studies as follows: the identification of time-variant parameters of a simply supported bridge when the vehicle passing, and the model updating of a six-story concrete frame structure with field test during the Yangbi earthquake excitation in Yunnan Province, China. The comparison results of the numerical studies show that the proposed method is superior to the conventional UKF algorithm for the time-variant parameter identification in convergence speed, accuracy and adaptability to the sampling frequency. The field test studies demonstrate that the proposed method can provide suggestions for solving practical problems.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.3
• /
• pp.268-272
• /
• 2009

In this paper, we propose a Gaussian Mixture Model (GMM) based incremental robust adaptation with a forgetting factor for the speaker verification. Speaker recognition system uses a speaker model adaptation method with small amounts of data in order to obtain a good performance. However, a conventional adaptation method has vulnerable to the outlier from the irregular utterance variations and the presence noise, which results in inaccurate speaker model. As time goes by, a rate in which new data are adapted to a model is reduced. The proposed algorithm uses an incremental robust adaptation in order to reduce effect of outlier and use forgetting factor in order to maintain adaptive rate of new data on GMM based speaker model. The incremental robust adaptation uses a method which registers small amount of data in a speaker recognition model and adapts a model to new data to be tested. Experimental results from the data set gathered over seven months show that the proposed algorithm is robust against outliers and maintains adaptive rate of new data.

• The Journal of the Acoustical Society of Korea
• /
• v.33 no.5
• /
• pp.326-334
• /
• 2014

This paper proposes a novel covariance matrix estimation scheme for minimum variance distortionless response (MVDR) beamforming. By accurately tracking direction-of-sound source arrival (DoA) information using audio-visual sensors, the covariance matrix is efficiently estimated by adopting a variable forgetting factor. The variable forgetting factor is determined by considering signal-to-interference ratio (SIR). Experimental results verify that the performance of the proposed method is superior to that of the conventional one in terms of interference/noise reduction and speech distortion.