• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.10 no.4
• /
• pp.157-162
• /
• 2010

The cart-pole balancing problem is a pseudo-standard benchmark problem from the field of control methods including genetic algorithms, artificial neural networks, and reinforcement learning. In this paper, we propose a novel approach by using online reinforcement learning(OREL) to solve this cart-pole balancing problem. The objective is to analyze the learning method of the OREL learning system in the cart-pole balancing problem. Through experiment, we can see that approximate faster the optimal value-function than Q-learning.

• The Journal of the Acoustical Society of Korea
• /
• v.19 no.5
• /
• pp.79-83
• /
• 2000

When passband ripple α/sub p/ and stopband attenuation α/sub s/ at the w/sub s/ where the stopband begins are specified in filter design, △α/sub s/ usually exceeds the specification by △α/sub s/ due to the necessity that the order n of the filter function be an integer. In this paper, we apply a trade-off method to remove the excess stopband attenuation △α/sub s/ for reducing the value of pole-Q and improving the characteristics of the Chebyshev filter function. We also apply the trade-off method of pole-Q reduction to the modified Chebyshev function, and then the 4 types of function have been analyzed to compare in frequency and time domain characteristics. The trade-off method reduces the pole-Q which influences the filter characteristics to maximum 49.6% without increase of the order n. Thus implies that they have the improved characteristics such as the reduced passband ripple and flatter delay characteristics as compared Chebyshev filter function before trade-off. And the unit step response shows shorter delay time and settling time in time domain performance.

• Proceedings of the KIPE Conference
• /
• 2007.07a
• /
• pp.328-330
• /
• 2007

This paper deals with the estimation method on the initial pole position of a z-axis permanent magnet linear synchronous motor(PMLSM) without magnetic pole sensors such as Hall sensors. The proposed method takes account of the z-axis conditions such as the gravitational force and also the load conditions. The algorithm consists of two steps. The first step is to estimate the initial q-axis approximately by monitoring the movements at predefined different test q-axes. The second step is to estimate the real q-axis as accurately as possible based on the results at three different test q-axes. Experimental results on the z-axis PMLSM show good estimation characteristics of the proposed method.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.22 no.6
• /
• pp.89-96
• /
• 1985

A pole-placement control of discrete, deterministic, single-input single-output nonmini-mum phase systems is considered using a model reference type approach. The proposed pole-placement controller is designed in the parameter form to make the transfer function of the controller equal to that of the reference model with only single variable polynomial S(q-1). The proposed adaptive pole-placement controllr is designed with the true system para-meters by applying the adaptation method to the proposed pole-placement controller.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.13 no.1
• /
• pp.41-45
• /
• 2008

This paper deals with the estimation method on the initial pole position of a z-axis permanent magnet linear synchronous motor(PMLSM) without magnetic pole sensors such as Hall sensors. The proposed method takes account of the gravitational force at z-axis and also the load conditions. The algorithm consists of two steps. The first step is to approximately estimate the initial q-axis by monitoring the movements due to the test current at predefined different test q-axes. The second step is to estimate the real q-axis as accurately as possible by using the outputs corresponding to torques due to the test current at three different test q-axes in order to avoid the effect of load mass variations. Experimental results on the z-axis PMLSM show good estimation characteristics of the proposed method irrespective of load mass conditions.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2013.04a
• /
• pp.321-326
• /
• 2013

This paper investigates the magnetic excitation force of a surface-mounted permanent magnet (SPM) motor according to the change of pole/slot combination. The characteristics of magnetic flux and radial magnetic force (RMF) due to pole/slot combination were analyzed by using magnetic circuit analysis. Also, the RMF of motors with the variable pole/slot combination was numerically simulated by using the finite element analysis to verify the result of the magnetic circuit analysis. This research shows that RMF ripple is reduced when the number of pole is smaller than the number of slot.

• Proceedings of the Korea Electromagnetic Engineering Society Conference
• /
• 2003.11a
• /
• pp.294-298
• /
• 2003

In this paper, the modeling and design of high-Q multilayer passives have been investigated, and multilayer diplexer for GSM/DCS applications has been designed and fabricated using the passives. To minimize the system, the configuration of a multilayer inductor has involved a square spiral structure. Modeling of a multilayer inductor was performed by the subsystems of distributed components, and using the modeling the optimal structures of the high-Q multilayer inductor could be designed by analyzing parasitics and couplings which affect their frequency characteristics. Multilayer diplexer for GSM/DCS application has been designed and fabricated using LTCC technology. LPF for GSM band had the passband insertion loss of less than 0.55 dB, the return loss of more than 12 dB, and the isolation level of more than 26 dB by locating attenuation pole at 1800 MHz. HPF for DCS band had the passband insertion loss of less than 0.82 dB, the return loss of more than 11 dB, and the isolation level of more than 38 dB by locating attenuation pole at 930 MHz.

• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2000.11a
• /
• pp.163-167
• /
• 2000

The eligibility is used to solve the credit-assignment problem which is one of important problems in reinforcement learning. Conventional eligibilities which are accumulating eligibility and replacing eligibility make ineffective use of rewards acquired in learning process. Because only an executed action in a visited state is learned by these eligibilities. Thus, we propose a new eligibility, called the weighted eligibility with which not only an executed action but also neighboring actions in a visited state are to be learned. The fuzzy Q-learning algorithm using proposed eligibility is applied to a cart-pole balancing problem, which shows improvement of learning speed.

• The Transactions of the Korean Institute of Power Electronics
• /
• v.10 no.2
• /
• pp.124-129
• /
• 2005

This paper presents an advanced method for an initial pole position estimation of a Permanent Magnet Linear Synchronous Motor(PMLSM) that has an accurate incremental encoder for servo applications but does not have Hall sensors as a magnetic pole sensor. By appropriately using the secant method as a numerical method the proposed algorithm finds either of two zero force positions and then the correct d-axis by applying a q-axis test current. It only requires the tuned current controller and the relative position information md so it can be simply applicable to a rotary PMSM. The experimental results show the validity of the proposed method, which has an excellent performance with respect to an accurate pole position estimation under the minimal moving distance(average of about 85㎛) during the estimation process.

• Proceedings of the KIEE Conference
• /
• 2007.10c
• /
• pp.47-49
• /
• 2007

This paper deals with the generating characteristic analysis of permanent magnet (PM) machines with multi-pole rotor and 3-phase stator windings considering losses such as copper loss, iron loss and mechanical loss. First, using d-q transformation, dynamic equations of PM machines are established. And then, characteristic equations for losses, power and efficiency are also derived. On the basis of d-q dynamic equations and characteristic equations, dynamic simulation algorithm is achieved by the MATLAB/SIMULINK. The simulation results are validated extensively by finite element (FE) analyses.