Search

In a waveguide or a transmission line, the power supplied by the source and the impedance can be adjusted by controlling the amount of the phase change of the signal when it propagates through the equivalent transmission line once the medium, the structure, and the load of the waveguide or the transmission line are chosen. By choosing an appropriate frequency, transmission line, and load, the negative power factor of the transmission line circuit is obtained.

Reliability of Core and IO FinFET is extensively investigated with various process steps at Fin, Source/ Drain, sacrificial Gate-Metal, and High-Pressure D2 Anneal. By modulating the process knobs, we quantified the effect of oxide traps (at bulk or interface) on reliability mechanisms of replacement metal gate (RMG). The results are summarized as a process-reliability optimization guideline.

Basic Reliability of IO FinFET is studied with various process options and extensive trap density investigation. It is observed that the NBTI margin on High Performance IO pFET is lowered by additional treatment to maintain the taller Fin profile. Trap density investigation with charge pumping identified the source of oxide traps in tall FinFET and the mechanism of NBTI degradation. Finally, after oxide trap reduction, FinFET process was improved for both reliability and performance.

CMOS Image Sensor(CIS) products need higher voltage device and better analog characteristics than conventional SOC & Logic products. This work presents newly developed 14nm FinFET process with 2.xV high voltage FinFET device characteristics showing excellent analog and low power digital characteristics comparing to 28nm planar process. Gm is improved by 30% and 67% in FinFET process for NMOS and PMOS, respectively. Rout characteristics increased by 40 times and 6 times over 28nm planar process. Interface state density(Nit) improved by more than 40% and flicker noise characteristics also improved by 64% and 42% for NMOS and PMOS, respectively. Digital logic Transistor ion-ioff performance improved by by 32% and by 211% for NMOS and PMOS, respectively compared to 28nm planar device and the chip power consumption of digital logic functional block reduced by 34% in real Si of 12M pixel product. 14nm FinFET process expected to improve power consumption by 42% in 144M pixel density.

This paper aims to propose a nonlinear attitude controller for a drone system controlled by the control momentum gyros (CMGs). The dynamic modeling and the control mechanism of the considering system are first investigated. After that, based on the assumption of the timescale separation, the proposed controller is designed using the feedback linearization methodology. To make the resultant controller follow the three-loop structure, specific forms of the desired error dynamics are utilized in the design step. The characteristics of the proposed method are investigated through numerical simulations. The results show that the proposed controller can provide a good tracking performance against model uncertainties as well as model nonlinearities.

A recently developed set of the equations for the phase dependent power and impedance of a transformer circuit shows that the power factor is adjustable. To adjust the impedance of the circuit to a desired value, the phase change of the flux should be controlled. In this study, the power equations of the transformer circuit are examined for high frequency cases, and the condition to maximize the difference in powers between the primary coil and the load is explored.

Recently, there was a report on a new set of equations for a transformer circuit which considers the attenuation and the phase change of the flux that occur as it propagates through the magnetic core. By controlling the amount of phase change, the impedance can be adjusted. To determine the effect of the added phase factor in the impedance, simple cases are examined in this study: the cases of zero power factor, zero resistance, negative resistance, and positive power difference.

New equations for a transformer circuit are derived when the attenuation and the phase change occurring as the magnetic flux propagates through the magnetic core are considered. The power and the impedance can be adjusted by controlling the phase change. The law of conservation of energy is clearly broken, and the power dissipated at the load can be larger than that provided by the source. An induction machine and an imperfect transformer circuit can be treated in a similar fashion.

The conventional equations for a transformer circuit does not take account of phase change of the flux in the magnetic core. The power and the impedance depend on the phase of the flux. The power factor can be adjusted so that the power is delivered more at the load than that supplied by the source, against the law of conservation of energy. The derived equations and experiment show that the law is violated and that the over-unity machine is plausible.

Partial reflector sensing head is implemented for active mode locking (AML) laser sensor using fiber Fabry-Perot interferometer (FPI). The length change of fiber FPI sensing head segments can be measured by sweeping the modulation frequency of AML laser.

Novel fiber laser with programmable wavelength output have been demonstrated for various biomedical imaging applications. Due to the filter-less wavelength tuning mechanism, flexible wavelength tuning and switching are possible with a high swept rate of up to MHz in the various wavelength bands.

Fan Out Wafer Level Package (FOWLP) and Panel Level Package (PLP) have been introduced as cost-performance-form factor alternatives conventional packages. This paper shows the enhanced features and performance of FOWLP for high frequency automotive ADAS applications, fingerprint sensors and medical applications. Small form factor (

The effect of Joule heating (JH) on electromigration (EM) was investigated using copper low-k interconnects. We found Black's empirical EM model works at a very wide stress conditions (I = 0.03 mA ∼ 3.5mA), where the temperature rise by JH ranges from 0°C to 240°C. EM modeling parameters were found to be 1.05eV and 1.4 for EM Ea and n, respectively. Extensive failure analysis and FEM simulations were carried out to understand the EM failure mode at various JH conditions and provide guidelines for the usage of JH-assisted EM test method and design-reliability rules.

We fabricated 8Mb 1T-1MTJ STT-MRAM macro embedded in 28nm CMOS logic platform by developing novel integration/stack/patterning technologies. MTJ memory cell array was successfully embedded into Cu backend without open fail and severe degradation of magnetic property. Advanced perpendicular MTJ stack using MgO/CoFeB was developed to show high TMR value of 180% after full integration. In addition, ion beam etching (IBE) process was optimized with power, angle, and pressure to reduce a short fail below 1 ppm. Through these novel technologies, we demonstrated highly functional and reliable 8Mb eMRAM macro having a wide sensing margin and strong retention property of 85 0C and 10yrs.

Nd-Fe-B permanent magnets have been widely used in many fields because of their outstanding magnetic properties[1]. However, Pr,Nd and Dy are now considered to be critical elements due to an increasing demand. Therefore, researchs on substitution of Pr-Nd by Ce or misch-metal alloys (MM) are attracting renewed attention because of their potential applications in permanent magnets[2] [3]. The intrinsic magnetic properties of La2Fe14B and Ce2Fe14B are far inferior to those of Nd2Fe14B, so it is hard to acheive high performance R-Fe-B magnets based on misch-metal which contains a majority of La and Ce elements through conventional methods.

Moisture impact on dielectric reliability was investigated with intentional moisture uptake and removal from porous low-k dielectric. In our tests, moisture can reduce TDDB lifetimes by 6 orders, while the kinetics is maintained (for both dry and wet samples). Interestingly, moisture from humid atmosphere does not increase the leakage current, but can reduce TDDB lifetimes significantly. So, the leakage current is not a good index to detect moisture-induced TDDB degradation. TDDB tests with various pre-bake conditions show moisture removal is not linear over bake temperatures, indicating multiple moisture states in low-k dielectric. Moisture diffusion, scaled from our test condition, shows moisture can diffuse few micro-meters in one second into low-k dielectric even at room temperature and impact dielectric reliability. To recover from the moisture damage, we recommend a high bake temperature (300°C∼350°C).

In this paper, a roller-cam clutch (RCC) mechanism was developed. The RCC can transfer bidirectional torque with high backdrivability, as well as increase actuation energy efficiency, in electrical exoskeleton robots. The developed mechanism was installed at the knee joint, which uses less metabolic energy, and unclutched during the swing phase, thereby functioning as a passive joint. The RCC aimed to increase actuation energy efficiency, and by generating zero impedance for users during the swing phase, to produce high backdrivability. To develop the mechanism, mathematical modeling of the RCC was conducted, with the design having more than three safety factors following optimization. Titanium (Ti-6AL-4V) material was used. Finally, modeling verification was done using ANSYS software.

High-speed frequency-domain terahertz (THz) coherence tomography is demonstrated using frequency sweeping of continuous-wave THz radiation and beam steering. For axial scanning, THz frequency sweeping with a kHz sweep rate and a THz sweep range is executed using THz photomixing with an optical beat source consisting of a wavelength-swept laser and a distributed feedback laser diode. During the frequency sweep, frequency-domain THz interferograms are measured using coherent homodyne detection employing signal averaging for noise reduction and used as axial-scan data via fast Fourier transform. Axial-scan data are acquired while scanning a transverse range of 100 × 100 mm(2) by use of a THz beam scanner with moving neither sample nor THz transmitter/receiver unit. It takes 100 s to acquire axial-scan data for 100 × 100 points with 5 averaged traces at a sweep rate of 1 kHz. THz tomographic images of a glass fiber reinforced polymer sample with artificial internal defects are presented, acquired using the tomography system.

Due to excellent magnetic properties, sintered Nd-Fe-B magnets have been used in various fields for many years [1, 2]. However, Nd-Fe-B magnets show a relatively low temperature stability of performance which is a main disadvantage for the application in some special cases. In order to improve the temperature stability of performance, Co, Dy and Tb elements have been added to the Nd-Fe-B sintered magnets [2, 3]. It should be pointed out that the addition of heavy rare earth (Tb, Dy) leads to a drastic reduction of the saturation magnetization. In the present work, Pr 34.4 Co 65.6 alloy powder was blended with Nd-Fe-B master alloy powder using dual-alloy method. The effects of Pr 34.4 Co 65.6 powder on the thermal stability of properties and the microstructure were investigated in details.

In this paper, an eye detection method is proposed based on the features of eye images generated in the gaze tracking environment. By configuring the corneal specular reflection detector, pupil detector, and eyelid detector in a cascade format, in actual applications reliable eye detection can be achieved even when an image of only one eye is captured because a part of the user moves out of camera range or when changes occur in the brightness and focus because the user moves back and forth. For the corneal specular reflection detector, a block setting method and a thresholding method are proposed to ensure the corneal specular reflection is not missed. Furthermore, a method of suppressing the effects of eyelashes and corneal specular reflection is proposed that allows more accurate determination of the pupil in the pupil detection process. The eyelid detector determines the presence/absence of an eyelid by using a vertical integral projection function. The eyes were detected reliably when the proposed methods were applied to the captured images in an actual gaze tracking system.

Too much information exist in ubiquitous environment, and therefore it is not easy to obtain the appropriately classified information from the available data set. Decision tree algorithm is useful in the field of data mining or machine learning system, as it is fast and deduces good result on the problem of classification. Sometimes, however, a decision tree may have leaf nodes which consist of only a few or noise data. The decisions made by those weak leaves will not be effective and therefore should be excluded in the decision process. This paper proposes a method using a classifier, UChoo, for solving a classification problem, and suggests an effective method of decision process involving only the important leaves and thereby excluding the noisy leaves. The experiment shows that this method is effective and reduces the erroneous decisions and can be applied when only important decisions should be made.

This paper proposes an algorithm to automatically detect synchronization error that can be occurred in the stereoscopic video system. Specially, this paper is aiming toward 3DTV broadcasting system that requires a real-time operation with high accuracy. Accordingly, we propose new similarity calculation algorithm by applying the spatiogram concept, which is proper to the sync error detection at stereoscopic video with real-time operation. In this paper, the spatiogram includes temporal information to increase an accuracy of sync error detection. This paper also introduces an implementation of stereoscopic video synchronizer for the purpose of practical use, which performs the sync error correction function. Finally, the proposed algorithms are evaluated and verified through an implemented of 3DTV experimental system.

This paper presents eye detection method for gaze tracking system using the features of eye and corneal reflection. The proposed method sequentially discards the regions that are expected not to be an eye by classifiers based on the features such as pupil intensity and appearance, corneal reflection intensity, and so on. The classifiers are designed with some empirical parameters, thus the proposed method does not need training process.

Classification is very imprtant research in knowledge discovery and machine learning. The decision tree is one of the well-known data mining methods. In general, a decision tree can be grown so as to have zero eeor on the training data set. If there is any noise in the data set or it does not completely cover the decision space, then over-fitting occurs and the tree needs to be pruned in order to accurately generalize the test data set. In this paper, we propose a pre-pruning method with majority and minority properties for the decision tree. It uses two kinds of qualifying criteria to consider whether the ration of the highest class of a subtree is the majority of the subtree or a minority of the overall tree. New measures for these are added to the classifier with the extended data expression. Experiments show that a clasifier using this pruning method can improve classification accuracy as well as reduce the size of the tree.

Kinematic calibration system for humanoid robot hands was developed using vision cameras and alignment stages. Analysis of vision images of visual marks attached on the finger skin could give accurate joint coordinates of all finger joints through calibration process. This system consists of image calibration, stage alignment, mark image analysis, error estimate of joint coordinates, and finally calibrated kinematics of robot hand.

VIATOP (VT), Aluminum Via connecting the bond pad and the Top Copper level, is the critical component for Bond Pad (BP) Electromigration (EM) in advanced technology nodes, where a smaller VT or its array is employed for maximum chip-scaling. In this study, we evaluated BP EM using various dimensions of VT, investigated the scaling effect, and proposed a BP EM model for current crowding & reservoir effect in the VT.

A 64Gb MLC NAND flash memory on 20nm CMOS technology has been developed. 135mm2 chip size is realized by 1-sided All-Bit-Line architecture and 128 cells in a string. 25MB/s program throughput with 2-bit/cell is achieved by reducing BL resistance and pump output loading using new BL control method and multi-split block decoder. This device also supports 400MB/s high speed interface.

This paper shows experimentally dual wavelength swept at multi band (O, C-Band) property with active mode locking (AML) method. Unlike conventional wavelength swept laser, AML wavelength swept laser does not require any wavelength selecting filter in the cavity. The cavity has two free spectral ranges (FSRs) depend on dual path configuration. This wavelength swept laser can be useful for wi de-band fiber-optic application.

Electromigration (EM) of Copper interconnects is studied under unipolar pulse current (UPC) at 1.25 kHz ∼ 1.25 MHz and bipolar pulse current (BPC) at 100 kHz ∼ 5.0 MHz. For symmetric BPC, where a time-averaged current density (J avg ) is 0, we found interconnect failures by EM-induced void formation. This means EM is not fully recovered during the opposite-polarity pulse current. The degree of EM recovery is found to be 0.7 ∼ 0.9, where 1 is for full-recovery. Also, the short lead EM effect is investigated under UPC. We found short lead EM lifetimes are modeled by peak current density (J peak ), instead of J avg . Also, we observed, during the off-time between low frequency pulse currents, mechanical back-stress may reduce the stress gradient built by UPC and increase the EM lifetimes of the short leads.

The parameter study on the grasping characteristics of the humanoid robot hand with a spherical four bar linkage as a finger joint. The humanoid robot hand has one thumb and three fingers, each of which has two spherical four bar linkages, respectively. The humanoid robot hand is 280mm long and less than 2.5kg by weight. The robot hand totally has 13 degree of freedom: one thumb has 4 degree of freedom: each finger has 3 degree of freedom. The spherical four bar linkage was specially designed to rotate like a human finger joint. The robot hand can perform the dexterous motion like a human hand but also the relatively powerful grasping. Two parameters for the grasping characteristics studied in this paper were a diameter of a cylinder to be grasped, the type of grasping. The contact force of the finger joints were calculated for four cases. Numerical analysis had been done in a position control mode by Recurdyn with Matlab/Simulink.

The scope of working motion is expressed in the angle of handling tool motion to move in required conditions with carrying a variety of tools such as hammer, grinder and roller. As this scope becomes larger and it is possible to do the handling work in complex shape and narrow block space. The driving axis of working motion is the number of axis of link type to construct the main axis of working assistant system. Generally, the working motion becomes much smoother as the number of this axis becomes larger. The kind of handling tools is generally classified into several categories to do the handling work in different handling method by using different handling tools such as hammer, grinder and roller. The number of driving axis of related motion is defined in relation with arm and wrist motion of worker when the operator in working suit is doing the handling work. The working velocity is defined as the velocity of handling tools to move in working path and it is closely related with the productivity of handling work. The working load is defined as the total force to be pulled and pushed at end of the working assistant system or to be hanged on it.

A polycrystalline silicon (pc-Si) film is produced by low-temperature LPCVD process on a poly silicon seed layer. The properties of such film are characterized and compared with the films on a glass substrate and a silicon wafer. The poly crystalline silicon seed layers are fabricated by aluminum-induced crystallization (AIC) process with a glass/Al/Al 2 O 3 /a-Si structure, in which the Al layer was deposited on the glass substrate by a DC magnetron sputter, and a-Si film by PECVD method, respectively. After the AIC process with 400°C for 150min, the poly silicon film has the average grain size of about 10μm. The poly silicon seed layer is thickened (i.e. the poly silicon film is deposited on the seed layer) by LPCVD process with 600°C for 60min. As results, it is observed that the quality of the poly silicon film on the seed layer is better than that on the glass and worse than that on the silicon wafer. But the deposited poly silicon absorber layer could not reflect the property of the poly silicon seed layer, due to the relatively low temperature process.

This paper presents the power assistance system of working habiliment type for painting worker in ship-building. The painting worker carrying several tools such as spray gun, brush and roller should be stayed to do painting job in closed working space or inside internal wall of ship large block and in long time. Specially, the power habiliment system is realized by developing the mechanism and control technology to reduce the working load to be transferred to the muscular system of worker. And also it is developed to realize the human friendly optimal working motion. This paper shows the design of multi-axis link mechanism to do painting job of a variety of motion. It shows the control technology of painting work system to reduce the effect of the weight of painting tool and wearable painting system and the working load of painting process such as the reaction force of spray gun.

As the NAND flash memory market grows rapidly due to various applications, such as USB devices, MP3 players, SSDs, cellular phones, and cameras, there is a requirement for high-density and low-cost devices. Two different approaches to meet these requirements are increasing data per cell and area scaling. 3b/cell or 4b/cell NAND flash memories were introduced as an effective way to lower cost [1,2]. However, these devices suffer from program performance degradation since tighter V th distribution is required. On the other hand, area scaling is a candidate to achieve low cost while maintaining high program performance even though there are several hurdles to overcome, such as FG coupling and charge retention [3]. As the cell size gets smaller, the V th distribution widens and the erase-write cycling margin is decreased by the floating-gate coupling ratio [4].

This paper deals with a sliding mode impedance control (SMIC) of end-effector of robot manipulator using a tracking control scheme and a real-time artificial intelligence algorithm based on a radial basis function neural networks(RBFNNs). To real-time estimate the design parameters of desired impedance model such as desired inertia, damping, and stiffness desired impedance, SMIC(sliding mode impedance control) with RBFNNs algorithm is proposed.

In this paper, a robust self-tuning controller for a velocity control of permanent magnet type stepping motor is proposed‥ The self-tuning theory provides a nonlinear modeling of a stepping motor drive system and could provide the controller with information regarding the reference variation, system noise, and parameter variation of the stepping motor through the on-line estimation. Thus, the proposed self-tuning controller can change the gains of the controller according to system conditions. The availability of the proposed controller is verified through experiment results.

With the rapid increase of the world’s population, the drastic changes in world’s food supply, and the limitation of land resources, the pressure for agriculture is greater than ever before. With the development of AI theories and technologies, the study in the classification of agriculture data becomes more advanced and intellectualized. This paper mainly discusses a specific decision tree classifier which is used to predict and classify agriculture data. It is capable of dealing with both complete data and incomplete data. So that it can be applied into the classification problem for all kinds of agriculture data sets. The experiments are designed to prove the advantage of the proposed algorithm.

Biological early warning systems(BEWS) for toxicity detection is developed. BEWS detect toxicity by tracking actions of the whole creatures in aqua.Collected data have created some noise, and therefore can decrease the performance of prediction. In the paper,a noise reduction is scheme is devised to improve the performance of the prediction. The experiment result shows the decreased error rate by the new method, and is far better than that of the former BEWS system.

In the real world concepts are often not stable but change with time. The underlying data distribution may change as well. The model built on old data will be necessarily updated. This problem is known as concept drift. Mining concept drifts is one of the most important fields in mining data stream. The paper presents a totally new gradual forgetting approach for mining concept-drift data stream. We firstly utilize UChoo to mine data stream with concept drift. UChoo defines a weight for each instance. The latest data which represents new data distribution has gradually higher weight than old data when time passing. The experiment result shows that the new method performs higher accuracy.

As radar signal environments become denser and more complex, the capability of high-speed and accurate signal analysis is required for ES (electronic warfare support) system to identify individual radar signals at real-time. In this paper, we propose the novel clustering algorithm of radar pulses to alleviate the load of signal analysis process and support reliable analysis. The proposed algorithm uses KDE (kernel density estimator) and its CDF (cumulative distribution function) to compose clusters considering the distribution characteristics of pulses. Simulation results show the good performance of the proposed clustering algorithm in clustering and classifying the emitters.

Classification, which is the task of assigning object (event) to one of several predefined classes, is an important problem in the field of machine learning and data mining. Boosting based technique use deterministic results of weak classifiers to compound them to a strong classifier while classifier can give class distribution as result. It involves losses of information. This problem can be solved by using probabilistic idea. In this case class distribution is used to compound classifiers. Probabilistic results compounding is presented in this paper applying to incremental learning. Probabilistic results expression is easily realized using extended data expression. Results of experiments show power when compare to Learn++, an incremental ensemble-based algorithm.

Biological early warning systems (BEWS) which detects toxicity by tracking the physiologic responses of the whole organisms has been developed in recent years. In the paper, we firstly apply the C4.5 decision tree Algorithm to the biological early warning systems and propose a new BEWS which can detect toxicity by classifying the flea's behavior data. The new system is meaningful to both biological field and data mining. The new classification application includes decision tree algorithm and feature selection. The new BEWS system consists of training and test process. The experiment results show that the proposed new bio-monitoring system is available for environmental toxicity detection.

In ubiquitous environment, too much information is generated from a lot of sensors, and people want to obtain the appropriately classified information from the information. Decision tree algorithm like C4.5 is much useful in the field of data mining or machine learning system. Because this is fast and deduces good result on the problem of classification. This paper proposes three methods using decision tree for solving a classification problem. First, this paper suggest about the extended data expression. Second, a classifier, UChoo, based on the extended data expression is described. Third, this paper is to describe about rule generation. The rules expressed in the newly suggested method have almost the same information contents as compared with the original data set. The information is gotten from the sensors becomes large amount of data as the ubiquitous computation environment develops, therefore it is impossible to keep all information in memory. However, using suggested method, this problem is solved smoothly without losing almost the information.

This paper presents a method of visualizing a multi-dimensional data set into a lower dimensional space, especially into a two-dimensional space, so that people can intuitively conceive the relations or the distance between the entities of the data. Kullback-Leibler divergence is used as the measure to evaluate the distance between the vectors of the probability distribution. The measured distance values are used to find the corresponding coordinates of the entities in a lower dimensional space. Here, the one variable stochastic simulated annealing (OVSSA) is employed as the optimization technique. Experiments show that this is a plausible way of visualizing the multi-dimensional data, letting people see the relations among the entities intuitively.

Biological early warning systems(BEWS) has been developed in recent years. BEWS detects toxicity by tracking the physiologic responses of the whole organisms. In the paper, we apply the classification technique to the biological early warning systems and propose a new BEWS which is meaningful to biological field. Meanwhile, how to select the features in such classification application is also a contribution of this paper. By using the fractal dimension theory, we define the input features which represent the organism characteristics in non-toxic or toxic environment. The experiment results show that the proposed new bio-monitoring system is effective for environmental toxicity detection.

Bootstrap technique has been successfully used in many signal processing systems for data classification. Some of such systems are based on ensemble-based algorithms. These algorithms use multiple classifiers, generally to improve classification performance: each classifier provides an alternative decision whose combination may provide a superior solution than the one provided by any single classifier. In this paper, UChooBoost, a new supervised learning ensemble-based algorithm for extended data, based on bootstrap technique, is proposed. UChoo classifier is used as weak learner. UChoo classifier gives extended results expression. These results are combined by using new weighted majority voting founded on extended result expression.