Search

Multivariate time-series prediction is a common task, but it often becomes challenging due to missing data caused by unreliable sensors and other issues. In fact, inaccurate imputation of missing values can degrade the downstream prediction performance, so it may be better not to rely on the estimated values of missing data. Furthermore, observed data may contain noise, so denoising them can be helpful for the main task at hand. In response, we propose a novel approach that can automatically utilize the optimal combination of the observed and the estimated values to generate not only complete, but also noise-reduced data by our own gating mechanism. We evaluate our model on incomplete real-world time-series datasets and achieved state-of-the-art performance. Moreover, we present in-depth studies using a carefully designed, synthetic multivariate time-series dataset to verify the effectiveness of the proposed model. The ablation studies and the experimental analysis of the proposed gating mechanism show that it works as an effective denoising and imputation method for time-series classification tasks.

International audience; Hazardous Noxious Substance (HNS) is transported entirely through large vessels, so there is always a potential risk of marine HNS spills. In the event of an HNS accident, it can cause enormous human and property damage, so prompt detection is required. However, there is a limit to human access by ship, we need to use remote sensing data. In this study, ground experiments using hyperspectral cameras were performed to construct a spectral library of HNS. We classified the HNS and non-HNS by applying the hyperspectral mixture algorithm, and presented the HNS detection probability for every pixel by calculating the spectrum-based abundance fraction. The results of this study are expected to be used to estimate the extent of HNS spill in the event of a marine HNS accident.

In this study, the effect of hyperspectral atmospheric correction on accuracy of the ship detection was investigated. We applied two atmospheric correction algorithms on airborne hyperspectral data and then detected ship features from reflectance corrected by each algorithms with an unsupervised target detection method. Both ATREM and FLAASH algorithms produce comparable atmospheric correction results but the algorithm which applied with further correction options optimized for coastal monitoring purpose reveals better detection results for ship detection. Compared to the result from ATREM, the result from FLAASH-corrected reflectance shows a distinct contrast between the ship and surrounding background seawater. Mean values of spectral angles between target and endmember spectra were 0.2388 (ATREM) and 0.2169 (FLAASH). This implies a further optimization of the atmospheric correction can improve the performance of target detection using airborne hyperspectral data, especially for maritime search purpose.

The SAR-derived wind fields presented the detailed structure of wind fields along the coastal areas, which had heretofore been unobtainable from scatterometer observation. Comparison of the retrieved SAR wind speeds with in-situ buoy wind measurements showed a small difference of less than 1 m/s, which implied that the results of SAR wind retrieval satisfied the limit of accuracy of satellite scatterometry. The retrieved SAR wind fields off the east coast of Korea during August 2007 showed the distinct patterns of low wind along the coastal region. The analysis of SAR wind fields with coincided SST images indicated that these spatial distinctions of SAR wind fields were associated with the upwelling events. Based on wind and SST data from satellite data and in-situ measurements, it was found that the changes in MABL stability caused by upwelling events dominantly generated the variation of SAR wind fields at the coastal regions.

As the automobile industry has recently adopted information technologies, the latter are being used to replace mechanical systems with electronically-controlled systems. Moreover, automobiles are evolving into smart cars or connected cars as they are connected to various IT devices and networks such as VANET (Vehicular Ad hoc NETwork). Although there were no concerns about the hacking of automobiles in the past, various security threats are now emerging as electronic systems are gradually filling up the interiors of many automobiles, which are in turn being connected to external networks. As such, researchers have begun studying smart car security, leading to the disclosure of security threats through the testing or development of various automobile security technologies. However, the security threats facing smart cars do not occur frequently and, practically speaking, it is unrealistic to attempt to cope with every possible security threat when considering such factors as performance, compatibility, and so forth. Moreover, the excessive application of security technology will increase the overall vehicle cost and lower the effectiveness of investment. Therefore, smart car security risks should be assessed and prioritized to establish efficient security measures. To that end, this study constructed a security risk assessment framework in a bid to establish efficient measures for smart car security. The proposed security risk assessment framework configured the assessment procedure based on the conventional security risk analysis model GMITS (ISO13335) and utilized 'attack tree analysis' to assess the threats and vulnerabilities.

This study is directed for the information security education of the elementary students. The purpose of this paper is to examine that information ethics and information security awareness affect information security practice of elementary students. 192 valid questionnaires from elementary students of South Korea were collected, and then these were analyzed using structural equation modelling. The results show that information ethics and information security awareness affect information security practice of elementary students. And the research findings confirmed that the information security vocabulary test was useful to assess security awareness levels for elementary students. This new approach can assist management to plan and evaluate information security education. This study presents herein implications and suggestions for further research.

This paper proposes a new integer motion estimation algorithm which is designed for hardware execution by modifying the conventional TZ search to allow parallel motion estimation of all PU partitions. The algorithm consists of three phases, the first, raster, and refinement searches. At beginning of each phase, the algorithm obtains the search points required by the original TZ search for all PU partitions in a CU. Then, all the search points are used for the estimation of the motion costs for all PUs and the best search points are selected for all PUs. Compare to the conventional TZ search algorithm, the proposed algorithm decreases BD-BR by 1.06%.

High-efficiency video coding (HEVC) is a new video coding standard which offers higher performance than previous video coding standards such as H.264/AVC. Merge mode is one of the new tools adopted in HEVC to improve the inter-frame coding efficiency. Merge mode saves the bits for the motion vector (MV) by sharing the MV with neighboring blocks. Merge mode estimation (MME) is the process of finding a merge mode candidate which achieves the highest compression efficiency. During MME, extensive computations are required due to the motion-compensation and the rate-distortion optimization. In this paper, the characteristics and the computational complexity in the MME process are analyzed. And then, the intrinsic hardware utilization degradation problem is discussed. To improve the utilization of the MME hardware, the hardware efficient MME scheme is proposed. Additionally, the hardware organization which independently performs MMEs for integer motion vector (IMV) and fractional motion vector (FMV) is proposed. Due to the proposed MME and the specially designed hardware organization for that, the MME hardware in this paper achieves high utilization of hardware resources and high throughput of MME processing. The proposed hardware processes 105,474 of 64×64 CTUs per second with a clock frequency of 366 MHz, and its gate counts are 320.6K.

This research develops an underwater construction robot to level rubbles on the seabed for port construction. The rubble leveling is carried out by an underwater robot equipped with imaging sonars and underwater cameras, LBL and gyroscope sensors. A virtual reality system is developed to visualize the robot"s figure and the topography over the working environment; hence, the robot is virtually tele-operated by an operator. This paper presents the robot’s system and control, and it describes the working process of the rubble leveling carried out by the robot. In addition, the performance of the robot is demonstrated through the experimental results in subsea. The working speed of the robot is faster than that of a human diver, and the robot can work longer than the diver who can work for a limited time to prevent submarine sickness. The robot is expected to have much higher efficiency in deep water where a human diver is unable to work.

Digital circuits used in sensor networks require longer battery life and do not demand a fast frequency of operation. Sub-threshold circuits for such applications are an attractive option. Three dimensional ICs (3DICs) on the other hand is an emerging technology which helps in miniaturization and reduction in interconnects, resulting in power saving and performance improvement. Several works on sub-threshold circuits and TSV based 3DICs have been studied independently but none have studied the impact of 3D stacking of sub-threshold circuits. We design and study an ultra-low power 2-tier 3D sub-threshold implementation of H.264 intra frame encoder that encodes video frames. The encoder consumes 0.73μW power at 16.13 KHz clock frequency for a typical application of encoding a Common Image Format (CIF) frame. The motivation is to assess the feasibility of the use of extreme low power video encoders in image sensor based sensor networks. Low power operation is highly beneficial to such unattended sensor networks by extending their battery life. Sub-threshold design helps us in this respect while 3D stacking minimizes footprint area, helps in off-chip to on-chip memory integration and improves timing performance.

This paper presents a novel architecture of intra 8×8 prediction for a high profile H.264 intra frame encoder. Intra 8×8 prediction consists of prediction and reconstruction phases and two phases are performed in a pipeline manner. In the 2-stage pipelined schedule, bubbles between prediction and reconstruction phases are the main bottleneck for fast 8×8 prediction. To reduce bubbles, prediction mode reordering and Inverse Discrete Cosine Transform (IDCT) direction control schemes are proposed. The proposed schemes reduce 115 cycles of bubbles to perform intra 8×8 prediction for a macroblock and achieve a 24.5% speed-up.

We reported development of the gas vapor detection system using a piezoelectric driven microcantilever. The microcantilevers were fabricated by surface and bulk micromachining process. In order to discriminate 2,4-dinitrotoluene (DNT) in the air, the fabricated microcantilevers surface was coated by biological receptor, peptide, which is able to detect unique chemical among a variety number of chemically similar material. It allowed that our cantilever sensor can detect the byproduct of explosive material, DNT, with concentration range of ppb level by overcoming the non-specificity of conventional chemical sensor with the polymer receptors.

This work presents the effects of the common-gate (CG) stage nonlinearity on the linearity of the cascode amplifier. Conventionally, the CG stage is assumed as an ideal current buffer in the CMOS cascode low noise amplifier (LNA) design, but the analysis shows that the CG stage limits the linearity of the cascode LNA as the gain increases, due to the finite output resistance of the CG stage and the parasitic capacitance at the interstage node. Therefore, the simple linearization of the CS stage has difficulties to enhance the gain of LNA as the operating frequency increases. To confirm the analysis, a 2 GHz CMOS LNA was designed and the gain and the nonlinearity were measured. The measurement results show that there exists optimum load impedance to achieve the maximum OIP 3 of the LNA as expected by the analysis.

In this paper, we propose a microfluidic chip to generate well-defined microcapsule. In order to make core-shell microcapsule, a simple design capable of 3-dimensional hydrodynamic focusing was embedded in the proposed microfluidic chip. Cell suspension is focused 3-dimensional hydrodynamic in microchannel by the hillock based geometries, and then the core-shell structure of alginate microcapsules with thin membrane can be clearly identified. The mouse embryonic carcinoma cell line, P19, was cultured in the capsules and spherical EB (more than 70 %) with strong compaction was successfully formed within 2 days.

A new technique based on buck-boost topology is proposed to equalize a series-connected battery stack in this paper. The proposed scheme transfers the energy from fully charged battery cell to the weakest charged battery using buck-boost operation. This operation maintains batteries at the same charge and voltage level. Unlike previous battery equalizing schemes, the new battery equalizer uses only one magnetic component, resulting low cost and small size. Experiment results are provided to verify the operation of the proposed battery equalizer.

This paper proposes a speed sensorless vector control system for a permanent magnet synchronous motor (PMSM) by using a parallel reduced-order extended Kalman filter. The proposed scheme can save computation time greatly and estimate rotor speed and position by using new algorithms according to driving regions. To verify the merit of the proposed scheme, a comparison of a new reduced-order EKF algorithm with a conventional EKF algorithm has been also made in terms of computation time. Simulation and experimental results validate the theoretical analysis and show the feasibility of the proposed estimation technique.

This paper investigates performances of an extended kalman filter (EKF)-based cost-effective drive system of a permanent-magnet synchronous motor (PMSM) for low-cost applications. In this paper, the EKF is used as a speed and position estimator and the motor is fed from a four-switch three-phase (FSTP) inverter instead of a conventional six-switch three-phase (SSTP) inverter. This configuration reduces the cost of the inverter, the switching losses, and the complexity of the control algorithms. The robustness of the proposed EKF-based FSTP inverter fed PMSM drive is verified by simulation results at different operating conditions. A comparison of the proposed FSTP inverter fed PMSM drive with a conventional SSTP inverter system is also made in terms of performance and error analysis. The proposed FSTP inverter fed PMSM drive is found quite acceptable considering its performance, cost reduction and other advantages.

A post-linearization technique for the differential CMOS LNA is presented. The proposed method uses an additional cross-coupled FET pair which generates the 3rd-order inter-modulation (IM3) current to cancel out the IM3 current of the differential amplifier while minimizing the degradation of noise figure and avoiding the gain reduction. This technique is applied to enhance the linearity of the differential CMOS LNA using 0.18-mum technology. The LNA achieved +10.2-dBm IIP3 with 13.7-dB gain and 1.68-dB NF at 2 GHz consuming 11.8-mA from a 1.8-V supply.

This paper proposes a novel fault detection algorithm in a brushless DC (BLDC) motor drive system. This proposed method is configured without the additional sensor for fault detection and identification. The fault detection and identification are achieved by a simple algorithm using the operating characteristic of the BLDC motor. The drive system after the fault identification is reconfigured by four-switch topology connecting a faulty leg to the middle point of DC-link using bidirectional switches. This proposed method can also be embedded into existing BLDC motor drive system as a subroutine without an excessive computational effort. The feasibility of a novel fault detection algorithm is validated in simulation and experiment.

A new technique based on buck-boost topology is proposed to equalize a series-connected battery stack in this paper. The proposed scheme transfers the energy from fully charged battery cell to the weakest charged battery using buck-boost operation. This operation maintains batteries at the same charge and voltage level. Unlike previous battery equalizing schemes, the new battery equalizer uses only one magnetic component, resulting low cost and small size. Experiment results are provided to verify the operation of the proposed battery equalizer.

This paper proposes a novel fault detection algorithm for a brushless DC (BLDC) motor drive system. This proposed method is configured without the additional sensor for fault detection and identification. The fault detection and identification are achieved by a simple algorithm using the operating characteristic of the BLDC motor. The drive system after the fault identification is reconfigured by four-switch topology connecting a faulty leg to the middle point of DC-link using bidirectional switches. This proposed method can also be embedded into existing BLDC motor drive systems as a subroutine without excessive computational effort. The feasibility of a novel fault detection algorithm is validated in simulation.

A novel asymmetric current-fed energy-recovery circuit (ERC) for a plasma display panel (PDP) is proposed. Two ERCs composed of single switch and different resonant inductors, ERC-Y and ERC-X, are used for both sides of the PDP, and slow discharging and fast charging times can be employed. The ERC-Y discharging Cp to zero employs the slow discharging time, which reduces a current stress on ERC-Y compared with ERC-X. The ERC-X charging Cp to Vs employs the fast charging time, which ensures a stable light emission. Furthermore, inductor currents are built up before the PDP is charged and discharged. These built-up inductor currents help to fully charge and discharge the PDP, achieve the zero voltage switching (ZVS) of main switches, reduce an electromagnetic interference (EMI) noises, and compensate for a large gas-discharge current. Therefore, it features the ZVS, low EMI noises, low current stress, and high energy-recovery capability. The operation of the proposed circuit, design considerations, and experimental results are presented to show the validity of the proposed circuit.

We propose a predictive location based relaying for hybrid cellular and ad-hoc systems that can provide interference minimization to uplink transmission. In the case of dense networks, the target system of this research, it may cause heavy overhead traffic to use established routing schemes for the uplink transmission. Therefore, we propose location information server (LIS) which managed the mobile's routing information with centralized scheme. To maintain the routing correctness, mobile must send location information within a relatively short time period. However, the tradeoff as the location-update period decreases is the increase in the overall cellular overhead. To overcome this problem, a predictive location-based routing scheme is proposed. Simulation results show that interference is greatly improved with very low cellular overheads.

In this paper, the fault tolerant system for BLDC motors has been proposed to maintain the control performance under switching device faults of inverter. The proposed fault tolerant system provides compensation for open?circuit faults and short-circuit faults in power converter. The fault identification is quickly achieved by simple algorithm using the characteristic of BLDC motor drives. The drive system after fault identification is reconfigured by the four-switch topology connecting a faulty leg to the middle point of DC-link using bidirectional switches. The feasibility of the proposed fault tolerant system is proved by simulation.

We propose a method to improve Al thin film in the chemical vapor deposition (CVD) using superior Al precursor, Aluminum borohydride trimethylamine (AlBT). The deposition conditions of AlBT, which can improve factors such as sheet resistance (Rs) and reflective index (R.I.) that are related with the morphology of the CVD-Al film are optimized. The CVD-Al film by this AlBT improves via profiles and via resistance properties. Superior Al precursor, AlBT, show low particle distribution owing to its thermal-stable chemical property. These effects have AlBT to be an excellent promising precursor of CVD-Al technology to metallization process of sub 80 nm device and to be strongly production-worthy in mass production system.

We have used the non-contact electrical property measurement system to characterize the electrical leakage property of high-K materials such as Al2O3 and HfO2 on a patterned wafer. The basic technology is to measure the surface voltage with micro Kelvin probe after the corona charge deposition on a measurement area. Because of the charge decay through a dielectric material, voltage-time spectra follow exponential time dependence that is the characteristic of leakage induced charge decay. We have measured the electrical leakage property of the storage capacitors on the direct cell area of DRAM device. The measured electrical leakage property can be classified according to the thickness of Al2O3 and HfO2. Since the electrical leakage property depends on a thickness of a dielectric material, voltage-time spectra show different shapes according to the HfO2 thickness. Using the technology, we can monitor the electrical leakage property of the storage capacitors of high-K materials on the direct cell area

This paper presents an underwater robot developed for harbor construction, which requires the robot to carry a heavy load with large workspace and high dexterity. The requirement can't be attained by a conventional serial typed robot but by a parallel typed robot. For underwater work, it is necessary to waterproof the robot and its sensors. Especially, a sensor waterproof is a main drawback to build the underwater robot. This leads us to develop a sensorless observer which gives the position information without any position sensor. We design a neural network to identify the displacement change of the hydraulic cylinder according to the voltage exerting at servo valve. This paper introduces the design strategy for the parallel typed robot and the development with hydraulic control and wireless communication. The robot is attached to the end of an excavator or a crane to carry a large stone to build a bank. This paper shows the experimental results and explains the performance of the robot.

A novel energy-recovery circuit (ERC) for a plasma display panel (PDP) is proposed in this paper. Two different ERCs, ERC-Y and ERC-X, are used for both sides of the PDP, and the slow falling and fast rising times can be employed. The ERC-Y discharging Cp to zero employs the slow falling time and is composed of an LC resonant circuit biased by 0.5V s , which reduces the conduction loss and a voltage drop across a parasitic resistance, and decreases a current stress. The ERC-X charging C p to V s employs the fast rising time and is composed of an LC resonant circuit biased by V s , which ensures the stable light emission, fully charges C p to Vs regardless of a parasitic resistance, and compensates for a large gas-discharge current. Therefore, it features the zero voltage switching (ZVS), low electromagnetic interference (EMI), low current stress, no severe voltage notch, and high energy-recovery capability. The operation of the proposed circuit and design considerations are discussed in detail, and experimental results are presented to show the validity of the proposed circuit.

In this paper, the fault tolerant system for BLDC motor has been proposed to maintain the control performance under the open fault of inverter. The fault identification method is proposed to two different types. The one is the method by the difference between reference and actual current. The other is the method using additional voltage sensors across lower legs of inverter. The reconfiguration scheme is achieved by the four-switch topology connecting a faulty leg to the middle point of DC-link using bidirectional switches. The proposed fault tolerant system quickly recovers the control performance by short detecting time and reconfiguration of system topology. Therefore, continuous free operation of the BLDC motor drive system after faults is available. The feasibility of the proposed fault tolerant system is proved by simulation.

In this paper, a motor control method is proposed to improve the performance of sensorless drive of BLDC motors. In the terminal voltage sensing method, which is a great portion of sensorless control, a precise rotor position cannot be obtained when excessive input is applied to the drive during synchronous operation mode. Especially in the transient state, the response characteristic decreases. To cope with this problem, the unknown input (back-EMF) is modeled as the additional state of system in this paper. Taking into account the disturbance adopted by the back-EMF, the observer can be obtained by an equation of the augmented system. An algorithm to detect the back-EMF of a BLDC motor using the state observer is constructed. As a result, a sensorless drive of BLDC motors that can strictly estimate rotor position and speed is proposed.

In this paper, a sensorless brushless DC (BLDC) motor drive using the fuzzy back-EMF observer is proposed to improve the performance of conventional sensorless drive methods. Most existing sensorless methods of the BLDC motor have low performance at transients or low speed range and occasionally require additional circuit. To cope with this problem, the estimation of a back-EMF using the fuzzy logic is suitable for high performance because the back-EMF of the BLDC motor has a trapezoidal shape. The proposed algorithm using fuzzy back-EMF observer can achieve robust control for the change of an external condition and continuously estimate speed of the rotor at transients as well as steady state. The robustness of the proposed algorithm is proved through the simulation and is compared with other sensorless drive methods.

We propose a new adaptive admission control scheme for multiuser OFDMA systems to allocate wireless resource with low complexity. In this adaptive admission control scheme, the number of users is estimated adaptively by using the required data rate, bit error rate (BER) and average channel gain of each user at frame by frame. The resource allocation complexity becomes lower since the proposed method estimates the approximate number of users to be supported and reduce the repetition of actual resource allocation. Simulation result shows that the proposed method has high accuracy when ACG (Adaptive Craving Greedy) is used as the resource allocation scheme.

We propose a predictive location based relaying for hybrid cellular and ad-hoc systems that can provide interference minimization to uplink transmission. Location Information Server (LIS) managed the mobile's routing information with centralized scheme. To maintain the routing correctness, mobile must send location information within a relatively short time period. However the tradeoff as the location-update period decreases is the increase in the overall cellular overhead. To overcome this problem, a predictive location-based routing scheme is proposed. Simulation results show that interference is greatly improved with very low cellular overheads.

This research develops the robot manipulator combined a parallel manipulator and wrist manipulator. The robot has a parallel manipulator with three linear actuators and a central axis for positional motions, i.e., x, y, and z and a wrist manipulator with two rotary actuators for rotational motions, i.e., yaw and pitch. The proposed parallel-serial typed manipulator which has between high stiffness of parallel manipulator and high flexibility of wrist manipulator is applied to in machining work such as cutting, milling, grilling, etc. This paper introduces the parallel-wrist typed robot as a machining center and an assembly robot for the complex shape. In order to achieve the optimal design, we define four parameters such as the radius of a base and a platform, the initial length of linear actuator, and the stroke of linear actuator. To determine the parameters, investigation is done i.e., kinematics, available workspace, and dexterity analysis. In results, we show four optimal design parameters of parallel manipulator meet a desired workspace including orientational work of wrist manipulator.

The organic light emitting diode (OLED) is proposed as a novel source in the microchip because it has ideal merits (various wavelengths, thin-film structure and overall emitting) for the integration. In this paper, we fabricated finger-type PIN photodiodes for the fluorescence detection and an advanced microchip with an integrated OLED. The finger-type in the diode design extended the depletion region and its internal resistance was reduced about 32% than that of a rectangular-type. The photodiodes had 0.1 nA leakage current and 8,720 sensitivity (I/sub light//I/sub dark/) at -1 V bias. The interference filter with 32 layers (SiO/sub 2/, TiO/sub 2/) was directly deposited on the photodiode. Deposited OLEDs increased the excitation efficiency and extremely simplified the integration process. The monolithically integrated microchip had a high sensitivity of 0.256 nA//spl mu/M and the LOD (limit of detection, S/N =2) of 0.01 /spl mu/M.

This paper presents a new speed control method using the acceleration feedforward compensation and using the disturbance torque estimate method. The proposed method improve transition response characteristic of system and has been a robust characteristic in the four-switch three-phase motor drive system in which the gain of speed controller cannot be made large enough. The simulation results prove the validity of the proposed method.

We introduced the novel porous low-k material with k~2.3 improved the physical properties for intermetal dielectric application in high speed memory device. Because this low-k material has both Si-H and Si-CH 3 bonds in its SiO2 film, it reduced the bowing and adhesion failure of P-HSQ and P-MSQ at via profile and metal wire. As this material was applied at device, the parasitic capacitance between metal wires was reduced till 20% compared with that of the HSQ

In this paper, a novel motor control method is proposed to improve the performance of sensorless drive of BLDC motors. In the terminal voltage sensing method, which is a great portion of sensorless control, a precise rotor position cannot be obtained when excessive input is applied to the drive during synchronous operation mode. Especially in the transient state, the response characteristic decreases. To cope with this problem, the unknown input (back-EMF) is modelled as the additional state of system in this paper. Taking into account the disturbance adopted by the back-EMF, the observer can be obtained by an equation of the augmented system. An algorithm to detect the back-EMF of a BLDC motor using the state observer is constructed. As a result, a novel sensorless drive of BLDC motors that can strictly estimate rotor position and speed is proposed.

This paper presents a novel robust speed control method based on the load torque observer of high-performance brushless DC (BLDC) motor. Recently, high-performance BLDC motor drives are widely used for variable speed drive systems of the industrial applications. In case of the control of robot arms and tracking applications with lower stiffness, we cannot design the speed controller gain to be very large from the viewpoint of the system stability. Thus, the feedforward compensator with disturbance torque observer was proposed. This method can improve the servo stiffness without increasing the speed controller gain. The enhanced speed control performance can be achieved and the speed response against the disturbance torque can be improved for high-performance BLDC motor drive systems in which the bandwidth of the speed controller cannot be made large enough. The load disturbance is compensated by detected load torque through the observer. The compensation current is made through q-axis current. The d-q transform of phase currents were possible by the Fourier series summation method. Consequently, the speed control for high-performance BLDC motor drives become improved. The simulation results for BLDC motor drive systems confirm the validity of the proposed method.

In this paper, a new zero voltage switching (ZVS) phase shift full bridge (PSFB) converter with low conduction losses is proposed. It is based on the PSFB converter with series-connected two transformers (SCTT), which features wide ZVS ranges. By adding a capacitor, the proposed converter overcomes the disadvantage of the based converter, such as the high circulating energy. Furthermore, the turn ratio of the transformers can be increased as well. Therefore, high efficiency of the proposed converter can be achieved. Operational principles and experimental results for a 100 W (5 V, 20 A) prototype are presented to validate the proposed converter.

Summary form only given, as follows. We describe a method for enhancing adhesion of Au on parylene-c coated surfaces For the ion-beam and plasma effect for enhancing surface energy, ion beam treatment and atmosphere plasma treatment were performed for the modification from hydrophobic to hydrophilic surface, piranha, PBST and BSA treatment were performed Tape-test analysis and piranha treatment were used for qualitative adhesive comparison of Au on parylene-c substrate. Contact angle measurements were performed before and after various treatments. Surface roughness and microstructures were observed using FESEM and AFM. From the results, the increase in adhesion is attributed to increased surface energy rather than hydraulic surface characteristics.

Summary form only given, as follows. Parylene-c (Pa-c) film is a pinhole-free barrier against moisture, chemical and biofluid. So it is candidate for low dielectric materials for the application of biosensors such as DNA and protein chip Pa-c and SiO/sub 2/ films are deposited on a Si substrate using a vapor deposition method and LPCVD, respectively. Au was then deposited on both substrates using the e-beam evaporation method. For improvement of Au adhesion on Pa-c and SiO/sub 2/ surfaces, surface modification of Pa-c and SiO/sub 2/ was performed using oxygen plasma. Pa-c and SiO/sub 2/ layers were exposed to an oxygen plasma using a reactive ion etcher at various power and times.

The BLDC (brushless DC) motor is characterized by linear torque to current and speed to voltage. It has low acoustic noise and fast dynamic response. Moreover, it has high power density with high proportion of torque to inertia in spite of small size drive. However, at high-speed operation, torque and speed response characteristic is deteriorated by the motor inductance components in stator windings. It is difficult that the BLDC motor is used to the industrial application requiring a wide range of operating speeds. Phase advance angle control method of different methods used to improve torque and speed response is used widely at high-speed operation. However, conventional phase advance angle control method was considered about forward motoring mode. Therefore, position information error of rotor that reverse motoring mode has phase lag angle of double comparing with forward motoring mode is occurred. Consequently, torque performance is destroyed significantly. In this paper, a control method using adjustment of the phase advance angle is proposed. This proposed method improves the torque and speed response characteristic by minimizing delay of current at high-speed operation. Simulated results prove the effectiveness of the proposed method through comparison with the conventional control method.