• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.142-145
• /
• 2003

Many control applications using Neural Network need a priori information about the objective system. But it is impossible to get exact information about the objective system in real world. To solve this problem, several control methods were proposed. Reinforcement learning control using neural network is one of them. Basically reinforcement learning control doesn't need a priori information of objective system. This method uses reinforcement signal from interaction of objective system and environment and observable states of objective system as input data. But many methods take too much time to apply to real-world. So we focus on faster learning to apply reinforcement learning control to real-world. Two data types are used for reinforcement learning. One is reinforcement signal data. It has only two fixed scalar values that are assigned for each success and fail state. The other is observable state data. There are infinitive states in real-world system. So the number of observable state data is also infinitive. This requires too much learning time for applying to real-world. So we try to reduce the number of observable states by classification of states with Self-Organizing Map. We also use neural dynamic programming for controller design. An inverted pendulum on the cart system is simulated. Failure signal is used for reinforcement signal. The failure signal occurs when the pendulum angle or cart position deviate from the defined control range. The control objective is to maintain the balanced pole and centered cart. And four states that is, position and velocity of cart, angle and angular velocity of pole are used for state signal. Learning controller is composed of serial connection of Self-Organizing Map and two Multi-layer Feed-Forward Neural Networks.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1996년도 Proceedings of the Korea Automatic Control Conference, 11th (KACC); Pohang, Korea; 24-26 Oct. 1996
• /
• pp.181-184
• /
• 1996

This paper deals with the design problem of model reference adaptive controllers for MIMO plants with unknown orders. A design scheme for an adaptive control system based on CGT theorem, which has hierarchical structures derived from backstepping strategies, is proposed for MIMO plants with unknown orders but with known relative MacMillan degrees(relative degrees for SISO plants). It is also shown that all the signals in the resulting control system are bounded, and that the asymptotic tracking is achieved in the case where reference inputs are step.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1997년도 한국자동제어학술회의논문집; 한국전력공사 서울연수원; 17-18 Oct. 1997
• /
• pp.223-226
• /
• 1997

Usually, the Evolutionary Algorithms(EAs) are considered more efficient for optimal, system design because EAs can provide higher opportunity for obtaining the global optimal solution. This paper presents a mechanism of co-evolution consists of the two genetic algorithms(GAs). This mechanism includes host populations and parasite populations. These two populations are closely related to each other, and the parasite populations plays an important role of searching for useful schema in host populations. Host population represented by feedforward neural network and the result of co-evolution we will find the optimal structure of the neural network. We used the genetic algorithm that search the structure of the feedforward neural network, and evolution strategies which train the weight of neuron, and optimize the net structure. The validity and effectiveness of the proposed method is exemplified on the stabilization and position control of the inverted-pendulum system.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.225-230
• /
• 2003

This paper proposes a desired trajectory and a control algorithm for a biped robot with curved soles. Firstly, we derived the desired trajectory from a model called the Moving Inverted Pendulum Mode (MIPM) of which a contact point of the foot is moving in the horizontal direction. A biped robot with curved soles is under-actuated system, because it has one contact point with the ground during the single supporting phase. Therefore, to solve the under-actuated problem, we changed control variables, used modified dynamic equations and used the computed torque control. The simulation results show that a biped robot with curved soles walks stably. Also, fast walking and natural motion of a biped robot can be implemented.

• 조명전기설비학회논문지
• /
• 제27권8호
• /
• pp.42-50
• /
• 2013

The reactionary responses to control human standing dynamics were estimated under the assumption that postural complexity mainly occurs in the mid-sagittal plane. During the experiment, the subject was exposed to continuous horizontal perturbation. The ankle and hip joint rotations of the subject mainly contributed to maintaining standing postural control. The designed mobile platform generated anterior/posterior (AP) motion. Non-predictive random translation was used as input for the system. The mean acceleration generated by the platform was measured as $0.44m/s^2$. The measured data were analyzed in the frequency domain by the coherence function and the frequency response function to estimate its dynamic responses. The significant correlation found between the input and output of the postural control system. The frequency response function revealed prominent resonant peaks within its frequency spectrum and magnitude. Subjects behaved as a non-rigid two link inverted pendulum. The analyzed data are consistent with the outcome hypothesized for this study.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2006년 학술대회 논문집 정보 및 제어부문
• /
• pp.108-110
• /
• 2006

This paper deals with a tracking problem for nonlinear systems using its T-S fuzzy model and internal model. We extend the internal model of linear systems to an internal model of T-S fuzzy systems to accompany with state error of zero. A sufficient condition of the existence of a tracking controller for T-S fuzzy systems is expressed by linear matrix inequalities. A system of inverted pendulum on cart is illustrated to verify our method.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1999년도 하계학술대회 논문집 B
• /
• pp.728-730
• /
• 1999

This paper suggests a new supervisory control input for improving response of nonlinear system with adaptive neural controller. The proposed control input is constructed using variable parameters which is adjusted by a separated control law. With the proposed algorithm, the information about the parameters are not required and error range is managed by designer in this scheme. The effectiveness of the proposed control scheme is demonstrated through the control of inverted pendulum.

• 전자공학회논문지C
• /
• 제34C권5호
• /
• pp.43-54
• /
• 1997

The objective of this paper is to present a controller-design method that can guarantee the global stability for nonlinear systems described by takagi-sugeno fuzzy models, and to apply the method to a typical nonlinear control problem. The presented method gives us a compensated fuzzy controller through the following major steps: First, if each local linear model of a given takagi-sugeno fuzzy system does not have the same input matrix, the method expands the system into the one with a method finds a takagi-sugeno fuzzy controller guaranteeing the global stability of the closed loop via solving relevant linear matrix inequalities. Compared to the conventional PDC (paralled distributed compensation) technique, the presented method has an advantage that trial-and-errors to check the global stability are not necessary. An illustrative simulation on the control of inverted pendulum is performed to demonstrate the applicability of the presented method, and its results show that a controller satisfying the global stability and robustness can be obtained by the method.

• 융합신호처리학회논문지
• /
• 제11권3호
• /
• pp.250-254
• /
• 2010

As the number of input and fuzzy set of a fuzzy system increase, the size of the rule base increases exponentially and becomes unmanageable (curse of dimensionality). In this paper, tree architectures of fuzzy controller (TAFC) is proposed to overcome the curse of dimensionality problem occurring in the design of fuzzy controller. TAFC is constructed with the aid of AND and OR fuzzy neurons. TAFC can guarantee reduced size of rule base with reasonable performance. For the development of TAFC, genetic algorithm constructs the binary tree structure by optimally selecting the nodes and leaves, and then random signal-based learning further refines the binary connections (two-step optimization). An inverted pendulum system is considered to verify the effectiveness of the proposed method by simulation.

• 대한전기학회논문지:시스템및제어부문D
• /
• 제54권4호
• /
• pp.228-234
• /
• 2005

In this paper, we proposed a self tuning adaptive fuzzy sliding control algorithms using gadient descent method to reduce chattering phenomenon which is viewed in variable control system. In design of FLC, fuzzy control rules are obtained from expert's experience and intuition and it is very difficult to obtain them. We proposed an adaptive algorithm which is automatically updated by consequence part parameter of control rules in order to reduce chattering phenomenon and simultaneously to satisfy the sliding mode condition. The proposed algorithm has the characteristics which are viewed in conventional VSC, e.g. insensitivity to a class of disturbance, parameter variations and uncertainties in the sliding mode. To demonstrate its performance, the proposed control algorithm is applied to an inverted pendulum system. The results show that both alleviation of chattering and performance are achieved.