• Journal of Electrical Engineering and information Science
• /
• v.2 no.6
• /
• pp.82-89
• /
• 1997

In this paper, an adaptive speed controller with an FNN(Feedforward Neural Network) is proposed for servo motor drives. Generally, the motor system has nonlinearities in friction, load disturbance and magnetic saturation. It is necessary to treat the nonlinearities for improving performance in servo control. The FNN can be applied to control and identify a nonlinear dynamical system by learning capability. In this study, at first, a robust speed controller is developed by Lyapunov stability theory. However, the control input has discontinuity which generates an inherent chattering. To solve the problem and to improve the performances, the FNN is introduced to convert the discontinuous input to continuous one in error boundary. The FNN is applied to identify the inverse dynamics of the motor and to control the motor using coordination of feedforward control combined with inverse motor dynamics identification. The proposed controller is developed for an SR motor which has highly nonlinear characteristics and it is compared with an MRAC(Model Reference Adaptive Controller). Experiments on an SR motor illustrate te validity of the proposed controller.

• 제어로봇시스템학회:학술대회논문집
• /
• 2004.08a
• /
• pp.1281-1285
• /
• 2004

There are numerous emotions in the human world. Human expresses and recognizes their emotion using various channels. The example is an eye, nose and mouse. Particularly, in the emotion recognition from facial expression they can perform the very flexible and robust emotion recognition because of utilization of various channels. Hybrid-feature extraction algorithm is based on this human process. It uses the geometrical feature extraction and the color distributed histogram. And then, through the independently parallel learning of the neural-network, input emotion is classified. Also, for the natural classification of the emotion, advancing two-dimensional emotion space is introduced and used in this paper. Advancing twodimensional emotion space performs a flexible and smooth classification of emotion.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.54 no.6
• /
• pp.417-422
• /
• 2005

This paper presents an EMG pattern classification method to identify motion commands for the control of the artificial arm by SOM-TVC(self organizing map - tracking Voronoi cell) based on neural network with a feature parameter. The eigenvalue is extracted as a feature parameter from the EMG signals and Voronoi cells is used to define each pattern boundary in the pattern recognition space. And a TVC algorithm is designed to track the movement of the Voronoi cell varying as the condition of additive noise. Results are presented to support the efficiency of the proposed SOM-TVC algorithm for EMG pattern recognition and compared with the conventional EDM and BPNN methods.

• Proceedings of the KIEE Conference
• /
• 2000.07b
• /
• pp.994-996
• /
• 2000

Switched reluctance motor generally operates in the magnetically saturated region because the saturation gives several benefits to its performance. This paper investigates the modelling and fuzzy tuning PI control of a nonlinear switched reluctance motor. The modelling is performed through neural network technique. Fuzzy auto-tuning PI control is designed for a robust performance in load and speed variations. Simulation and experimental results indicate better performances compared with simple PI control.

• ETRI Journal
• /
• v.41 no.2
• /
• pp.235-241
• /
• 2019

In this paper, we propose a rank-weighted reconstruction feature to improve the robustness of a feed-forward deep neural network (FFDNN)-based acoustic model. In the FFDNN-based acoustic model, an input feature is constructed by vectorizing a submatrix that is created by slicing the feature vectors of frames within a context window. In this type of feature construction, the appropriate context window size is important because it determines the amount of trivial or discriminative information, such as redundancy, or temporal context of the input features. However, we ascertained whether a single parameter is sufficiently able to control the quantity of information. Therefore, we investigated the input feature construction from the perspectives of rank and nullity, and proposed a rank-weighted reconstruction feature herein, that allows for the retention of speech information components and the reduction in trivial components. The proposed method was evaluated in the TIMIT phone recognition and Wall Street Journal (WSJ) domains. The proposed method reduced the phone error rate of the TIMIT domain from 18.4% to 18.0%, and the word error rate of the WSJ domain from 4.70% to 4.43%.

• Journal of Information Processing Systems
• /
• v.19 no.6
• /
• pp.730-744
• /
• 2023

The topic of this paper is the recognition of human activities using egocentric vision, particularly captured by body-worn cameras, which could be helpful for video surveillance, automatic search and video indexing. This being the case, it could also be helpful in assistance to elderly and frail persons for revolutionizing and improving their lives. The process throws up the task of human activities recognition remaining problematic, because of the important variations, where it is realized through the use of an external device, similar to a robot, as a personal assistant. The inferred information is used both online to assist the person, and offline to support the personal assistant. With our proposed method being robust against the various factors of variability problem in action executions, the major purpose of this paper is to perform an efficient and simple recognition method from egocentric camera data only using convolutional neural network and deep learning. In terms of accuracy improvement, simulation results outperform the current state of the art by a significant margin of 61% when using egocentric camera data only, more than 44% when using egocentric camera and several stationary cameras data and more than 12% when using both inertial measurement unit (IMU) and egocentric camera data.

• Proceedings of the Korean Institute of Industrial Safety Conference
• /
• 2003.10a
• /
• pp.3-16
• /
• 2003

After a not-widely-known fact is revealed that Japan is a leading country in production and use of industrial unmanned helicopters, a kind of UAV. The voice command system and the autonomous flight control system with a variety of control algorithms including neural network, robust and adaptive control that have been developed in collaboration between Kyoto University and Yamaha Motor Co., and funded by the Ministry of Education and Science of Japan are described in some detail. Both already-proven and promising future applications of the autonomous unmanned helicopters are given.

• Proceedings of the KIEE Conference
• /
• 2007.07a
• /
• pp.1644-1645
• /
• 2007

This paper proposes an adaptive sliding mode control method for trajectory tracking of nonholonomic mobile robots with model uncertainties and external disturbances. The kinematic model represented by polar coordinates are considered to design a robust control system. Wavelet neural networks (WNNs) are employed to approximate arbitrary model uncertainties in dynamics of the mobile robot. From the Lyapunov stability theory, we derive tuning algorithms for all weights of WNNs and prove that all signals of an adaptive closed-loop system are uniformly ultimately bounded.

• Proceedings of the KIEE Conference
• /
• 1995.07a
• /
• pp.68-70
• /
• 1995

The controller for magnetically suspended vehicles considering the speed characteristics is presented. Generally, the attraction force of magnet is determined by input current and air gap. However, when MAGLEV runs at high speed, induced eddy currents in the rail decrease the attraction force. Thus control characteristics of MAGLEV become deteriorated. Therefore, the variation of attraction force according to speed must be considered. Thus we analyzed the speed characteristics of the magnet by FEM. Also, we study on the control characteristics according to speeds, and design the controller considering the decline of levitation force using Neural Network.

• 한국태양에너지학회:학술대회논문집
• /
• 2008.04a
• /
• pp.61-67
• /
• 2008

In this paper proposed the solar tracking system to use a fuzzy based on PC in order to increase an output of the PV array. The solar tracking system operated two DC motors driving by signal of photo sensor. The control of dual axes is not an easy task due to nonlinear dynamics and unavailability of the parameters. Recently, artificial intelligent control of the fuzzy control, neural-network and genetic algorithm etc. have been studied. The fuzzy control made a nonlinear dynamics to well perform and had a robust and highly efficient characteristic about a parameter variable as well as a nonlinear characteristic. Hence the fuzzy control was used to perform the tracking system after comparing with error values of setting-up. nonlinear altitude and azimuth. In this paper designed a fuzzy controller for improving output of PV array and evaluated comparison with efficient of conventional PI controller. The data which were obtained by experiment were able to show a validity of the proposed controller.