• Journal of the Korea Computer Graphics Society
• /
• v.27 no.1
• /
• pp.9-17
• /
• 2021

In this paper, we present a simple and fast supervised learning framework based on model predictive control so as to learn motion controllers for a physic-based character to track given example motions. The proposed framework is composed of two components: training data generation and offline learning. Given an example motion, the former component stochastically controls the character motion with an optimal controller while repeatedly updating the controller for tracking the example motion through model predictive control over a time window from the current state of the character to a near future state. The repeated update of the optimal controller and the stochastic control make it possible to effectively explore various states that the character may have while mimicking the example motion and collect useful training data for supervised learning. Once all the training data is generated, the latter component normalizes the data to remove the disparity for magnitude and units inherent in the data and trains an artificial neural network with a simple architecture for a controller. The experimental results for walking and running motions demonstrate how effectively and fast the proposed framework produces physics-based motion controllers.

• The Transactions of the Korean Institute of Electrical Engineers
• /
• v.45 no.1
• /
• pp.113-122
• /
• 1996

As integrity of semiconductor device is increased, accurate and efficient modeling and recipe generation of semiconductor fabrication procsses are necessary. Among the major semiconductor manufacturing processes, dry etc- hing process using gas plasma and accelerated ion is widely used. The process involves a variety of the chemical and physical effects of gas and accelerated ions. Despite the increased popularity, the complex internal characteristics made efficient modeling difficult. Because of difficulty to determine the control input for the desired output, the recipe generation depends largely on experiences of the experts with several trial and error presently. In this paper, the optimal control of the etching is carried out in the following two phases. First, the optimal neural network models for etching process are developed with genetic algorithm utilizing the input and output data obtained by experiments. In the second phase, search for optimal control inputs in performed by means of using the optimal neural network developed together with genetic algorithm. The results of study indicate that the predictive capabilities of the neural network models are superior to that of the statistical models which have been widely utilized in the semiconductor factory lines. Search for optimal control inputs using genetic algorithm is proved to be efficient by experiments. (author). refs., figs., tabs.

• Journal of the Korean Society for Precision Engineering
• /
• v.26 no.5
• /
• pp.120-127
• /
• 2009

An electric furnace, inside which desired temperatures are kept constant by generating heat, is known to be a difficult system to control and model exactly because system parameters and response delay time vary as the temperature and position are changed. In this study the heating system of ceramic drying furnaces with time-varying parameters is mathematically modeled as a second order system and control parameters are estimated by using a RIV (Recursive Instrumental-Variable) method. A modified bang-bang control with magnitude tuning is proposed in the time optimal temperature control of ceramic drying electric furnaces and its performance is experimentally verified. It is proven that temperature tracking of adaptive time optimal control using a second order model is more stable than the GPCEW (Generalized Predictive Control with Exponential Weight) and rapidly settles down by pre-estimation of the system parameters.

• Journal of Korea Technical Association of The Pulp and Paper Industry
• /
• v.37 no.1 s.109
• /
• pp.67-72
• /
• 2005

In the grade change operations inputs and outputs are highly correlated and application of conventional linear feedback control methods such as PID schemes might lead to poor control performance. In this study the neural networks model for the grade change operation is trained by using bilinear terms which can represent non-linear characteristics of grade change operations. The inverse model of the grade change operation is obtained from training and the optimal input variables are computed from the trained neural networks as well. The proposed bilinear inverse model predictive control scheme was found out to showlittle discrepancy between simulated outputs and setpoints.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 1998.10a
• /
• pp.324-329
• /
• 1998

In this paper, neural networks predictive PID (NNPPID) control system is proposed to reduce the vibration of structure. NNPPID control system is made up predictor, controller, and self-tuner to yield the optimal parameters of controller. The neural networks predictor forecasts the future outputs based on present input and output of structure. The controller is PID type whose parameters are yielded by neural networks self tuning algorithm. Computer simulations show displacements of multi-story structures applied to NNPPID system about environmental load-wind forces and earthquakes.

• Industrial Engineering and Management Systems
• /
• v.1 no.1
• /
• pp.1-9
• /
• 2002

Among the state-space description of discrete vent systems, the max-plus algebra is known as one of the effective approach. This paper proposes a model predictive control (MPC) design method based on the max-plus algebra. Several studies related to these topics have been done so far under the constraints that system parameters are constant. However, in practical systems such as production systems, it is common and sometimes inevitable that system parameters vary by each event. Therefore, it is of worth to design a new MPC controller taking account of adjustable system parameters. In this paper, we formulate system parameters as adjustable ones, and they are solved by a linear programing method. Since MPC determines optimal control input considering future reference signals, the controller can be more robust and the operation cost can be reduced. Finally, the proposed method is applied to a production system with three machines, and the effectiveness of the proposed method is verified through a numerical simulation.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.13 no.5
• /
• pp.2414-2419
• /
• 2012

This paper presents an optimal operation control method for energy saving in the water reuse pumping system. A predictive horizon switching strategy is proposed to implement an optimal operation control and a linear programming (LP) algorithm is used to solve optimal problems in each time step. Energy costs are calculated for electricity on both TOU in the light, heavy, and maximum load time period and peak charges. The optimal operation in water reuse pumping systems is determined to reduce the TOU and peak costs. The simulation results show a power energy saving for water reuse pumping systems and power stability improvement.

• 제어로봇시스템학회:학술대회논문집
• /
• 1988.10a
• /
• pp.693-696
• /
• 1988

This paper is investigated on the improvement for steam temperature control at Boryung coal-fired drum boiler type thermal power plant. The steam temperatur control has been mainly operated by the feedback controllers. Automatic controllers are bounded and difficult. Because boiler system is nonlinear and the system time delay is very large. Optimal regulators including predictive feedforward and differentiate control are synthesized and some improved output results are presented.

• 제어로봇시스템학회:학술대회논문집
• /
• 1997.10a
• /
• pp.325-328
• /
• 1997

In this paper the modeling and control of a twin roll strip caster is investigated. Mathematical models for the strip casting process are obtained by analyzing five critical areas such that the molten steel level in the pool, solidification process, roll separating force and torque, roll dynamics including hydraulic actuators, and roll drive system. A two-level control strategy is proposed. At lower level, three local subsystems are independently feedback-controlled by suitable local controllers which perform well to the behaviors of each subsystem. They are a variable structure control of the molten steel level in the pool, an adaptive predictive control of the roll gap which is directly related to the strip thickness, and an $H^{\infty}$ control of the roll drive system. At higher level, all reference signals to the lower level subsystems are generated by an optimal controller in the perspective of regulating the strip thickness and roll separating force. Simulations are provided..

• International Journal of Control, Automation, and Systems
• /
• v.1 no.4
• /
• pp.412-430
• /
• 2003

In many practical problems in signal processing and control, the signal values are often restricted to belong to a finite number of levels. These questions are generally referred to as "finite alphabet" problems. There are many applications of this class of problems including: on-off control, optimal audio quantization, design of finite impulse response filters having quantized coefficients, equalization of digital communication channels subject to intersymbol interference, and control over networked communication channels. This paper will explain how this diverse class of problems can be formulated as optimization problems having finite alphabet constraints. Methods for solving these problems will be described and it will be shown that a semi-closed form solution exists. Special cases of the result include well known practical algorithms such as optimal noise shaping quantizers in audio signal processing and decision feedback equalizers in digital communication. Associated stability questions will also be addressed and several real world applications will be presented.