• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.22 no.7
• /
• pp.993-1000
• /
• 2018

This paper proposes a low-dropout voltage regulator(LDO) using self-cascode structure. The self-cascode structure was optimized by adjusting the channel length of the source-side MOSFET and applying a forward voltage to the body of the drain-side MOSFET. The self-cascode of the input differential stage of the error amplifier is optimized to give higher transconductance, but the self-cascode of the output stage is optimized to give higher output resistance, The proposed LDO using self-cascode structure was designed by a $0.18{\mu}m$ CMOS technology and simulated using SPECTRE. The load regulation of the proposed LDO regulator was 0.03V/A, whereas that of the conventional LDO was 0.29V/A. The line regulation of the proposed LDO regulator was 2.23mV/V, which is approximately three times improvement compared to that of the conventional LDO. The transient response of the proposed LDO regulator was 625ns, which is 346ns faster than that of the conventional LDO.

• The Transactions of the Korea Information Processing Society
• /
• v.6 no.3
• /
• pp.692-698
• /
• 1999

In this paper, we proposed the channel assignment scheme for the CDMA(Code Division Multiple Access) cellular system using power allocation concept. Also, the performance of the proposed scheme was analyzed and it was compared with the channel assignment scheme using the fixed power allocation method in the CDMA cellular system. The proposed scheme allocates the power adaptively in according to the traffic loads and the traffic distribution pattern of neighbor cells in the forward link. We found that total call blocking probability (Pr) is more dependent on blocking probability($P_B$) than outage probability (Po) under physical number of channels ($C_{th}$)=30. Pr(Call Blocking Probability) is dependent on $P_B$(Blocking Probability) and Po(Outage Probability) at the same ratio under $C_{th}$=32, in which case P$P_{TA}$(blocking probability for the adaptive power allocation) is greater than $P_{TF}$(blocking probability for the fixed power allocation) about 6%.

• Journal of Satellite, Information and Communications
• /
• v.6 no.2
• /
• pp.20-25
• /
• 2011

The authors have proposed novel TDD-CDMA systems with Pre-Rake transmit diversity schemes (system 1 and system 2) using multiple transmit antennas in [2] and have also evaluated the system performance through the theoretical analysis and computer simulation. However, the performance of system 2 which transmit a signal using all antennas has not been evaluated for multi-user environment. Therefore in this paper, we analyze the performance of system 2 for multi-user environment and compare the performance with that of the already proposed system 1 which chooses only one antenna. From the numerical results, it is found that system 2 outperforms system 1 as the number of users increases while system 1 outperforms system 2 at a small number of users. Therefore in order to achieve the best system performance, the Pre-Rake transmit diversity type should be selected at the base station according to the number of users.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.10C
• /
• pp.1433-1443
• /
• 2004

The power line modem proposed in this paper transmits the remote control signal using CSK(Code Shift Keying) and DS/SS method. The CSK technique provides the increased capacity of transmission and robustness towards noise. Besides, the DS/SS technique provides protection against narrow-band Gaussian interference and multi-path interference. The modem supports full-duplex communication using FDD(Frequency Division Duplex) and the modem structure for forward link is same with that for reverse link. To switch each sub-controlled unit smoothly, 4/$\pi$-DQPSK is adopted for noncoherent demodulation. The PN code for spreading spectrum seues to divide each group which consists of sub-controlled units and Walsh code is used for the M-ary CSK technique. Each block is designed and verified with TMS320C5402 DSP. We show the superiority of the proposed method by analyzing numerically the system performance for the factors of the DS/SS and CSK method ullder additive white Gaussian noise and PBI.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2003.11a
• /
• pp.19-22
• /
• 2003

In this paper, a new small size Lateral Trench Electrode Power MOSFET with local doping is proposed. This new structure is based on the conventional lateral power MOSFET. The entire electrodes of proposed device are placed in trench oxide. The forward blocking voltage of the proposed device is improved by 3.3 times with that of the conventional lateral power MOSFET. The forward blocking voltage of proposed device is about 500V. At the same size, a increase of the forward blocking voltage of about 3.3 times relative to the conventional lateral power MOSFET is observed by using TMA-MEDICI which is used for analyzing device characteristics. Because the electrodes of the proposed device are formed in trench oxide respectively, the electric field in the device are crowded to trench oxide. And because of the structure which has a narrow drain doping width, the punch through breakdown can be occurred in higher voltage than that of conventional lateral power MOSFET. We observed that the characteristics of the proposed device was improved by using TMA-MEDICI and that the fabrication of the proposed device is possible by using TMA-TSUPREM4.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2004.07a
• /
• pp.292-295
• /
• 2004

The SIT was introduced by Nishizawa. in 1972. When compared with high-voltage, power bipolar junction transistors, SITs have several advantages as power switching devices. They have a higher input impedance than do bipolar transistors and a negative temperature coefficient for the drain current that prevents thermal runaway, thus allowing the coupling of many devices in parallel to increase the current handling capability. Furthermore, the SIT is majority carrier device with a higher inherent switching speed because of the absence of minority carrier recombination, which limits the speed of bipolar transistors. This also eliminates the stringent lifetime control requirements that are essential during the fabrication of high-speed bipolar transistors. This results in a much larger safe operating area(SOA) in comparison to bipolar transistors. In this paper, vertical SIT structures are proposed to improve their electrical characteristics including the blocking voltage. Besides, the two dimensional numerical simulations were carried out using ISE-TCAD to verify the validity of the device and examine the electrical characteristics. A trench gate region oxide power SIT device is proposed to improve forward blocking characteristics. The proposed devices have superior electrical characteristics when compared to conventional device. Consequently, the fabrication of trench oxide power SIT with superior stability and electrical characteristics is simplified.

• Journal of the Korean Vacuum Society
• /
• v.14 no.4
• /
• pp.229-237
• /
• 2005

The degradation phenomena in SiGe hetero-junction bipolar transistors(SiGe HBTs) induced by bias stress are investigated in this review. If SiGe HBTs are stressed over a specific time interval, the device parameters deviate from their nominal values due to the internal changes in the devices. Reverse-bias stress on emitter-base(EB) junctions causes base current increase and current gain decrease because carriers accelerated by the electrical field generate recombination centers. When forward-bias current stress is conducted at an ambient temperature above $140^{\circ}C$ , hot carriers produced by Auger recombination or avalanche multiplication induce current gain fluctuation. Mixed-mode stressing, where high emitter current and high collector-base voltage are simultaneously applied to the device, provokes base current rise as EB reverse-bias stressing does.

• Journal of Korea Society of Industrial Information Systems
• /
• v.24 no.5
• /
• pp.53-64
• /
• 2019

The term 're-use' here signifies the re-use of end-of-life products without changing their form after they have been thoroughly inspected and cleaned. In the re-use network model, the distributor determines the product order quantity on the network through which new products are received from the suppliers or products are supplied to the customers through re-use of the recovered products. In this paper, we propose a reusable network model for reusable products that considers the total logistics cost from the forward logistics to the reverse logistics. We also propose a reusable network model that considers the processing and disposal costs for reuse in an optimal inventory management environment. The authors employe Genetic Algorithm (GA), which is one of the optimization techniques, to verify the validity of the proposed model. And in order to investigate the effect of the parameters on the solution, the priority-based GA (priGA) under three different parameters and the modified Hybrid GA (mhGA), in which parameters are adjusted for each generation, were applied to four examples with varying sizes in the simulation.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.30 no.6
• /
• pp.1053-1065
• /
• 2020

With the recent development of hardware performance and artificial intelligence technology, sophisticated fake videos that are difficult to distinguish with the human's eye are increasing. Face synthesis technology using artificial intelligence is called Deepfake, and anyone with a little programming skill and deep learning knowledge can produce sophisticated fake videos using Deepfake. A number of indiscriminate fake videos has been increased significantly, which may lead to problems such as privacy violations, fake news and fraud. Therefore, it is necessary to detect fake video clips that cannot be discriminated by a human eyes. Thus, in this paper, we propose a deep-fake detection model applied with Bidirectional Convolution LSTM and Attention Module. Unlike LSTM, which considers only the forward sequential procedure, the model proposed in this paper uses the reverse order procedure. The Attention Module is used with a Convolutional neural network model to use the characteristics of each frame for extraction. Experiments have shown that the model proposed has 93.5% accuracy and AUC is up to 50% higher than the results of pre-existing studies.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.22 no.4
• /
• pp.95-102
• /
• 2022

This paper proposed a linear time algorithm for a three-partition problem(TPP) in which a polynomial time algorithm is not known as NP-complete. This paper proposes a backtracking method that improves the problems of not being able to obtain a solution of the MM method using the sum of max-min values and third numbers, which are known polynomial algorithms in the past. In addition, the problem of MM applying the backtracking method was improved. The proposed algorithm partition the descending ordered set S into three and assigned to the forward, backward, and best-fit allocation method with maximum margin, and found an optimal solution for 50.00%, which is 5 out of 10 data in initial allocation phase. The remaining five data also showed performance to find the optimal solution by exchanging numbers between surplus boxes and shortage boxes at least once and up to seven times. The proposed algorithm that performs simple allocation and exchange optimization with less O(k) linear time performance complexity than the three-partition m=n/3 data, and it was shown that there could be a polynomial time algorithm in which TPP is a P-problem, not NP-complete.