• Transactions of the Korean Society of Mechanical Engineers
• /
• v.19 no.9
• /
• pp.2105-2113
• /
• 1995

The compressive tests under impact conditions were performed to establish a design guide for impact damage tolerance. The composition of layup was selected for the real cases of composite aircraft structure. The energy level of visible of visible damage threshold was determined as 7 Joules. It was found that the normalized bending stiffnesses in the direction of closely fixed boundary affected the area of damage. Graphite/epoxy used in the tests exhibited 60% reduction in compression strength at the energy level of visible damage threshold. Wet-conditioned specimens represented 9% reduction in residual compressive strength in comparison with room temperature ambient specimens. In this study, a design factor of 2.1 was proposed for the low velocity impact damage.

• Journal of information and communication convergence engineering
• /
• v.13 no.2
• /
• pp.113-122
• /
• 2015

Feature subset selection is as a pre-processing step in learning algorithms. In this paper, we propose an efficient algorithm, ModifiedFAST, for feature subset selection. This algorithm is suitable for text datasets, and uses the concept of information gain to remove irrelevant and redundant features. A new optimal value of the threshold for symmetric uncertainty, used to identify relevant features, is found. The thresholds used by previous feature selection algorithms such as FAST, Relief, and CFS were not optimal. It has been proven that the threshold value greatly affects the percentage of selected features and the classification accuracy. A new performance unified metric that combines accuracy and the number of features selected has been proposed and applied in the proposed algorithm. It was experimentally shown that the percentage of selected features obtained by the proposed algorithm was lower than that obtained using existing algorithms in most of the datasets. The effectiveness of our algorithm on the optimal threshold was statistically validated with other algorithms.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.17 no.10 s.113
• /
• pp.985-992
• /
• 2006

The classic routing algorithms try to minimize the number of hops to reach a destination. In our DPC(Distributed Power Control) this hop-by-hop power level selection is also used to select the path guaranteeing low energy consumption. The main goal is to realize a dynamic protocol able to be aware of mobility and to automatically set security threshold in order to get the best performance in every situation. In conclusion, the simulation DPC has improved energy gain of the AMRIS routing algorithm at the narrow area, and ODMRP routing algorithm has improved energy gain at the wide area.

• ETRI Journal
• /
• v.27 no.6
• /
• pp.685-690
• /
• 2005

DC and RF characteristics of $0.15{\mu}m$ GaAs power metamorphic high electron mobility transistors (MHEMT) have been investigated. The $0.15{\mu}m{\times}100{\mu}m$ MHEMT device shows a drain saturation current of 480 mA/mm, an extrinsic transconductance of 830 mS/mm, and a threshold voltage of -0.65 V. Uniformities of the threshold voltage and the maximum extrinsic transconductance across a 4-inch wafer were 8.3% and 5.1%, respectively. The obtained cut-off frequency and maximum frequency of oscillation are 141 GHz and 243 GHz, respectively. The $8{\times}50{\mu}m$ MHEMT device shows 33.2% power-added efficiency, an 18.1 dB power gain, and a 28.2 mW output power. A very low minimum noise figure of 0.79 dB and an associated gain of 10.56 dB at 26 GHz are obtained for the power MHEMT with an indium content of 53% in the InGaAs channel. This excellent noise characteristic is attributed to the drastic reduction of gate resistance by the T-shaped gate with a wide head and improved device performance. This power MHEMT technology can be used toward 77 GHz band applications.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.11 no.5
• /
• pp.365-370
• /
• 1998

We studied a electrical properties in GaAs/AlGaAs/InGaAs pseudomorphic high electron mobility transistors(PHEMT s) recessed by electron cyclotron resonance(ECR) plasma and wet etching. Using the $NH_4OH$ solution, a nonvolatile AlF$_3$layer formed on AlGaAs surface after selective gate recess is effectively eliminated. Also, we controlled threshold voltage($V_th$) using $H_3PO_4$ etchant. We have fabricated a device with 540 mS/mm maximum transconductance and -0.2 V threshold voltage by using $NH_4OH$ and $H_3PO_4$dip after ECR gate recessing. In a 2-finger GaAs PHEMT with a gate length of 0.2$\mu m$ and width of 100 $\mu m$, a current gain of 15 dB at 10 GHz and a maximum cutoff frequency of 58.9 GHz have been obtained from the measurement of current gain as a function of frequency at 12mA $I_{dss}$ and 2 V souce-drain voltage.

• Proceedings of the Korean Institute of Electrical and Electronic Material Engineers Conference
• /
• 2003.07b
• /
• pp.1094-1097
• /
• 2003

In this paper, the cavity-length-dependent spectral and temporal characteristics of a V-groove AlGaAs-GaAs quantum wire (QWR) laser at each subband were investigated. At short cavity lasers less than $300{\mu}m$, a discrete wavelength switching from the n=1 to the n=2 subband occurred due to the increased threshold gain, resulting from the increased cavity loss. Using the characteristic of the wavelength shift from n=1 to the n=2 subband with shortening the cavity length, ultrafast lasing behaviors under gain switching at the n=1 and the n=2 subband transition were demonstrated and compared.

• Journal of the Korean Institute of Telematics and Electronics A
• /
• v.32A no.2
• /
• pp.48-56
• /
• 1995

The operating characteristics, such as, the threshold gain, lasing frequency, and longitudinal intensity profile, etc., of 1.55$\mu$m DFB laser diode with index and/or gain grating structures and with one side AR-coated mirror have been analyzed. From this analysis, the optimum design parameters have been shown that ${\Delta}{\Omega}$ (the phase difference between index grating and gain grating) is 0 or ${\pi}$, (xL)$_{r}$=1~3 and (xL)$_{i}$=0.5~0.9. It has been also shown that the modal selectivity and intensity uniformity of the DFB lasers with .DELTA..OMEGA.=0 are ~1.2 times better thatn those of the DFB lasers with ${\Delta}{\Omega}$= ${\pi}$.

• Journal of IKEEE
• /
• v.14 no.2
• /
• pp.90-97
• /
• 2010

The high efficiency power management IC(PMIC) with DTMOS(Dynamic Threshold voltage MOSFET) switching device and DTMOS Error Amplifier is presented in this paper. PMIC is controlled with PWM control method in order to have high power efficiency at high current level. Dynamic Threshold voltage CMOS(DT-CMOS) with low on-resistance is designed to decrease conduction loss. The control parts in Buck converter, that is, PWM control circuits consist of a saw-tooth generator, a band-gap reference circuit, an DT-CMOS error amplifier and a comparator circuit as a block. the proposed DT-CMOS Error Amplifier has 72dB DC gain and 83.5deg phase margin. also Error Amplifier that use DTMOS more than CMOS showed power consumption decrease of about 30%. DC-DC converter, based on Voltage-mode PWM control circuits and low on-resistance switching device is achieved the high efficiency near 96% at 100mA output current. And DC-DC converter is designed with Low Drop Out regulator(LDO regulator) in stand-by mode which fewer than 1mA for high efficiency.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.11A
• /
• pp.1113-1119
• /
• 2006

For a multi-user OFDM system in mobile channels which requires low-complexity in adaptive resource allocations, resource allocation algorithm using multi-threshold is proposed. The allocation scheme, which is performed by the multi-threshold values in descending order, considers only subcarriers over each threshold level. Moreover, some subcarriers with the lowest channel gain can be· removed in the present threshold level within the constraint of satisfaction of the required data rate, in order to allocate them to the other users when the allocation process of next threshold is executed. As a result, the proposed bandwidth and subcarrier algorithm has better system performances than the conventional allocation schemes in terms of required power and processing time, which is expected as a technique that improves the spectral efficiency of OFDM systems in a mobile environment.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.24 no.6
• /
• pp.628-633
• /
• 2014

This paper proposes a performance improvement method of the conventional double talk detection method which can operate before convergence of the echo canceller. The proposed method estimates the coefficients of the linear predictive coding(LPC) filter by using the primary input signal. The time-varying threshold for double talk detection is determined based on the LPC filter gain of the primary input signal level. The proposed method can reduce not only false detection rate which means wrong detection of single talk as double talk but also double talk detection delay. Computer simulation was performed using a long-term real speech signals. It is shown that the proposed method improves the conventional method in terms of lowering the false detection rate and shortening the detection delay.