Session Schedule

F1J. Transmission Lines and Sensor Applications

Location: 3F Room J
Date & Time: 09:30-10:30, Thursday, 16 October 2008
Chair: Yuji Ohashi, Assistant Professor

F1J-1
Fiber sensor for refractive measurement based on surface plasmon resonance
Lin Ma, Yuji Matsuura

-- Abstract --

We propose a fiber sensor for high resolution refractive measurementbased on surface plasmon resonance (SPR) by using a tapered Braggfiber. This unique structure allows core effective refractive index ofthe fiber to be continuously adjusted at any wavelength to achieve thedesired SPR. The fiber sensor structure is optimized and a steepincrease of the transmission loss due to the SPR is obtained when thedevice surrounded in an environment with a designed refractive index.The prototype fiber sensor with a sensing length of 10 mm has beenfabricated and an increase of transmission loss of 7 dB is obtained.

F1J-2
A Study on Measurement of Permittivity of Liquid Using a Cylindrical Cavity
Eiki Kushizaki, Jerdvisanop Chakarothai, Qiang Chen, Kunio Sawaya, Makoto Suzuki

-- Abstract --

Method of permittivity of measurement for liquid by using a coaxial probe is often practically used. This method assumes that structure of sample is infinite, so there is no guarantee that the result is correct or not in case there is plenty of sample. Besides, it is needed that sample size should be small. Therefore, in this research, the effect of method of permittivity measurement for liquids by using cylindrical cavity whose volume is small is studied.

F1J-3
High-Accurate Measurement of Refractive Indices of Liquid Crystals Using Multiple-Interference Tri-Incidences Method
Yuji Ohno, Takahiro Ishinabe, Tetsuya Miyashita, Tatsuo Uchida

-- Abstract --

A new method for measuring the ordinary and extraordinary refractive indices of liquid crystal materials was devised. The refractive indices can be obtained from the numerical fitting between measured values and calculated values considering multiple interference of the amplitude ratios and the phase differences at three incidences. Moreover, we developed the new Jones matrix method considering interference in a LC cell. The validity of our new method (Multiple Interference Tri-Incidences (MITI) method) was confirmed from experimental results which the refractive indices did not depend on conditions of LC cell. The MITI method is effective for the design of high-quality LCDs.

F1J-4
Measurement of Electromagnetic Near-field Around a Transmission Line and Effect of Position of a Connector Contact Failure
Yu-ichi Hayashi, Hideaki Sone

-- Abstract --

A transmission line created by cables adjoined by connectorsisinfluenced by noise from connectors with contact failure, andsuchnoise degrades communication quality. The authors used a model ofaconnector with increased contact resistance in a coaxialtransmissionline and measured the electromagnetic near-field around acable whilechanging positions of the model. In this paper, the resultshows thatthe radiated electromagnetic field has no relationship withtheposition along the cable of a connector with increasedcontactresistance, when the contact condition of connector,contactresistance value, measurement position, and length of atransmissionline are constant.

Go back to the Top.

F1K. Invited Lecture I

Location: 3F Room K
Date & Time: 09:30-10:30, Thursday, 16 October 2008
Chair: Qiang Chen, Associate Professor

Delivering Spectral Efficient and Robust Wireless Solutions - An Overview of Wireless Communications Research at Institute for Infocomm Research (I²R)
Sumei Sun, Tjeng Thiang Tjhung
Institute for Infocomm Research

-- Abstract --

We will give an overview on the wireless communication researchactivities at the Institute for Infocomm Research (I²R),Singapore. In particular, we will present our research findings inpoint-to-point multiple-input multiple-output (MIMO) orthogonalfrequency division multiplexing (OFDM) to achieve high spectralefficiency and robust performance, in robust cognitive radio (CR)spectrum sensing to achieve high spectrum utilization efficiency, andin high-speed ultra-wideband (UWB) communication systems. We will alsointroduce a low-complexity multi-antenna OFDM relay system in whichonly pre-fast-Fourier-transform (pre-FFT) time-domain processing isrequired at the relay to realize second-order Alamouti space-timeblock coded diversity.

Go back to the Top.

F1L. Invited Lecture II

Location: 3F Room L
Date & Time: 09:30-10:30, Thursday, 16 October 2008
Chair: Yukio Iwaya, Associate Professor

Overview of Algorithms for Swarm Intelligence
Jeng-Shyang Pan, Pei-Wei Tsai, Shu-Chuan Chu
Department of Electronic Engineering, National Kaohsiung University of Applied Sciences

-- Abstract --

This paper reviews some popular algorithms in the field of swarm intelligence for problems of optimization. The overview and experiments of Particle Swarm Optimization (PSO), Ant Colony System (ACS), Cat Swarm Optimization (CSO), the parallel versions of ACS, PSO and CSO and the Bacterial-GA Foraging are given. In addition, some comparisons are made between these algorithms. To inspect more details of these algorithms, the related information can be found in the reference.

Go back to the Top.

F2J. Plasma Science and Surface Science

Location: 3F Room J
Date & Time: 10:45-11:45, Thursday, 16 October 2008
Chair: Kousaku Miyake, Assistant Professor

F2J-1
Effects of Plasma Parameters on Synthesis and Properties of Nitrogen Atom Encapsulated Fullerene Using an RF Plasma
Jamal Uddin Ahamed, S. Miyanaga, T. Kaneko, R. Hatakeyama

-- Abstract --

Atomic nitrogen is encapsulated inside a sublimated fullerene (C60) molecule in a radio frequency (RF) discharge plasma, which is confirmed by an electron spin resonance analysis. It is found that plasma parameters have a remarkable affect on the synthesis of the nitrogen atom encapsulated fullerene (N@C60). The relatively high purity of N@C60 of about 0.08% is achieved using an electron beam superimposed RF plasma whose parameters are easily controlled. The purpose of this research is to elucidate the synthesis mechanism of N@C60 in order to improve the purity of the synthesized N@C60, and to investigate the optical properties of the high purity N@C60.

F2J-2
Control of Electron Temperature Gradient in Electron Cyclotron Resonance Plasmas
Mohammad Maksudur Rahman, S. Tamura, T. Kaneko, R. Hatakeyama

-- Abstract --

At present it is an important issue to find out the best plasma condition for material processing such as atom or molecule encapsulated fullerenes which have potential applications in organic semiconductor or quantum computation. In the present work control of electron temperature gradient in an electron cyclotron resonance (ECR) plasma is examined. The microwave produced ECR plasma passes through different sized mesh grids to an experimental region where low-temperature thermionic electrons are superimposed with the ECR plasma. It is found that the electron temperature can be controlled in a spatially localized area when low-temperature thermionic electrons are superimposed upon the high-temperature ECR plasma and the negative bias voltage of the grid is increased.

F2J-3
Formation of Gas-Atom Encapsulated Silicon Fullerenes Using Silicon Plasmas
Masahiro Yabuno, Toshiro Kaneko, Rikizo Hatakeyama

-- Abstract --

Formation of nanosized spherical shaped silicon (Si) clusters, i.e., Si fullerene is desired to be investigated, because Si is the most important semiconductor material in minute electronics. It is found pure Si clusters are formed using a Si plasma generated by an electron beam gun in the absence of an additional gas. In the case of including an argon (Ar) gas in the Si plasma, on the other hand, not only the pure Si clusters but also atom-doped Si clusters are found to be produced. In addition, it is detected by a X-ray photoemission spectroscope that the doped Si clusters are mainly composed of Si and Ar. Based on these results, structure of the Si clusters is considered to be a spherical shape containing the Ar atom at the center, namely, an Ar encapsulated Si fullerene appears to be successfully formed.

F2J-4
Gold catalyzing growth of single walled carbon nanotubes using chemical vapor deposition method
Zohreh Ghorannevis

-- Abstract --

Single-Walled carbon nanotubes (SWNTs) are attracting much interest due to their unique electronic, mechanical properties and the potential applications in various fields. We have reported growth of SWNTs by thermal chemical vapor deposition (TCVD) from ethanol and also poorly crystallized carbon nanotubes using plasma chemical vapor deposition (PCVD) from methane feedstock over gold catalyst. Effects of hydrogen on growing carbon nanotubes during growth and pretreatment were investigated by Raman spectroscopy measurements in order to find the optimum condition to grow SWNTs over gold catalyst. We observed that increasing the hydrogen concentration during growth while keeping the ethanol vapor pressure constant reduces the yield of carbon nanotubes, which might be attributed to etching the active species. On the other hand, introducing hydrogen gas during pretreatment can increase the yield of SWNTs, which can be explained by increasing the catalytic activity of gold nanoparticles by hydrogen pretreatment.

Go back to the Top.

F2K. Wireless Communications I

Location: 3F Room K
Date & Time: 10:45-11:45, Thursday, 16 October 2008
Chair: Hiroyasu Sato, Assistant Professor

F2K-1
Single-carrier Transmission with Joint Tomlinson-Harashima Precoding/Frequency-domain Pre-equalization
Kazuki Takeda, Hiromichi Tomeba, Fumiyuki Adachi

-- Abstract --

Frequency-domain equalization (FDE) significantly improves the performance of single-carrier (SC) transmission to suppress the inter-symbol interference (ISI) in a severe frequency-selective channel. However, the residual ISI after FDE limits the performance improvement. Recently, we proposed a joint use of Tomlinson-Harashima precoding (THP) and FDE (joint THP/FDE) to suppress the residual ISI. However, joint THP/FDE requires the channel state information (CSI) at both a transmitter and a receiver. In this paper, we replace FDE by frequency-domain pre-equalization (pre-FDE) and propose joint THP/pre-FDE that requires the CSI at the transmitter only, while providing almost the same BER performance as joint THP/FDE.

F2K-2
Throughput Performance of OFDM/TDM Using MMSE-FDE in a Multipath Channel
Haris Gacanin, Fumiyuki Adachi

-- Abstract --

Orthogonal frequency division multiplexing (OFDM) with hybridautomatic repeat request (HARQ) using incremental redundancy strategyachieves a good throughput performance over a wireless channel.However, during thefirst transmission uncoded information packet is transmitted andconsequently, the channel frequency-selectivity cannot be exploited.In this work, to increase the throughput of OFDM, we apply OFDMcombined with time division multiplexing (OFDM/TDM) using minimum meansquare error frequency domain equalization (MMSE-FDE). It is shown, bycomputersimulation, that OFDM/TDM using MMSE-FDE achieves a higher throughputthan the conventional OFDM due to frequency diversity gain during thefirst transmission.

F2K-3
Capacity Analysis for Downlink Cellular MIMO Systems
Wei Peng, Fumiyuki Adachi

-- Abstract --

As the demand for reliable high speed data transmissionincreases, the capacity of downlink cellular multipleinputmultiple-output (MIMO) systems is of much interest.However, the capacity analysis regarding the frequency reusefactor (FRF) is rarely reported. In this paper, theoretical analysisfor both ergodic and outage capacities for cellular MIMOsystems are presented. It is shown by the numerical resultsthat FRF 1 can achieve the maximal capacity for best situationwhile FRF 3 can achieve the maximal capacity for theworst situation.

F2K-4
Adaptive inter-relay data exchange in cooperative wireless relay network
Salma Ait Fares, Fumiyuki Adachi, Eisuke Kudoh

-- Abstract --

We proposed recently a novel cooperative relay network based on data exchange between relays to enhance the transmit signals from relays to the destination. In this inter-relay process occurred during an additional time-slot, only the relay possessing the highest SNR is engaged into exchanging data by forwarding its received signal to the other relays. However, when the SNR values of relays are almost equal, this inter-relay process will not be required and result in throughput reduction. In this paper, we adopt an adaptive time-slot to our proposed method to increase the throughput performance when the inter-relay process is not needed.

Go back to the Top.

F2L. Invited Talk

Location: 3F Room L
Date & Time: 10:45-11:45, Thursday, 16 October 2008
Chair: Yukio Iwaya, Associate Professor

F2L-1
An Application: Cat Swarm Optimization for Opposing DC Components Watermarking
Pei-Wei Tsai, Jeng-Shyang Pan, Bin-Yih Liao
Department of Electronic Engineering, National Kaohsiung University of Applied Sciences

-- Abstract --

The Cat Swarm Optimization (CSO) is a new proposed algorithm for solving optimization problems in 2006. According to the experimental results, the CSO is useful for solving optimization problems. On the other hand, the DC components watermarking is a watermarking scheme on the transformed domain, which has the high robustness and good transparency. The opposing method for the DC components watermarking is proposed in 2008. An application of CSO for opposing DC components watermarking is proposed in this paper. The experimental result indicates that the CSO can efficiently select a suitable set of parameters for the opposing method.

F2L-2
Efficient Scalable Distributed Video Coding Based on Residual SW-SPIHT
Anhong Wang, Yao Zhao, Jeng-Shyang Pan
Institute of Information Science, Beijing Jiaotong University, Taiyuan University of Science and Technology

-- Abstract --

In this paper, a novel scheme for scalable distributed video (SDVC) isintroduced, which dealing with quality scalabilities. Morespecifically, efficient H.264/AVC intra-frame coding is used to obtaina base quality layer, the residual between the base layer and theoriginal video is encoded by distributed video coding (DVC) withSW-SPIHT to generate the enhancement layer. The side information isgenerated by the residual between the base layer and the frameinterpolated by motion estimation. Since the residual coding exploitsthe similarity between the base layer and enhancement layer,experimental results show this SDVC approach is more efficient thanthe referenced but with similar encoding computation.

Go back to the Top.

F3J. Invited Lecture III

Location: 3F Room J
Date & Time: 13:00-14:00, Thursday, 16 October 2008
Chair: Tetsuya Miyashita, Associate Professor

Eco Displays
Han-Ping David Shieh
National Chiao Tung University

-- Abstract --

High efficiency, low material cost, green are among key features for eco-displays. Liquid Crystal Display (LCD) is the most popular display nowadays, yet the optical efficiency is only around 10%, most of power is turned into undesired heat. Among the components to generate color image on LCD, color filter and polarizer are the two components of the lowest efficiency: of less than 30% and 50%. Field sequential color liquid crystal displays (FSC-LCDs) use temporal domain, instead of commonly used spatial domain to form color images. By using red (R), green (G), and blue (B) LED backlight array and temporal color sequential scheme in conjunction with novel fast switching LC, no color filters are needed. Consequently, optical throughput can be increased substantially, power consumption can be lowered, and image quality of factor of three (no R, G, B sub-pixels for coloring as conventional displays) in resolution and color rendering can be enhanced. In this talk, the suppression of the serious color break-up (CBU) artifact and improvement of image quality, the human visual model for CBU and novel image driving algorithm for an Eco Display will be presented.

Go back to the Top.

F3K. Invited Lecture IV

Location: 3F Room K
Date & Time: 13:00-14:00, Thursday, 16 October 2008
Chair: Fumiyuki Adachi, Professor

Wireless Communications with Multiple Antennas
Qi-Tu Zhang
City University of Hong Kong

-- Abstract --

Multiple-input multiple-output (MIMO) antenna technique is a powerfulmeans to exploit huge potential capacity in multipath fading and thus,has been widely adopted in wireless communications. In this talk, wepresent a concise overview of various relevant issues of MIMOtechnology, aiming to fully understand its methodology,transceiverdesign, and applications. Some of our recent results in MIMO,multihop, and distributed wireless systems are also presented forillustration.

Go back to the Top.

F3L. Image Processing I

Location: 3F Room L
Date & Time: 13:00-14:00, Thursday, 16 October 2008
Chair: Koichi Ito, Assistant Professor

F3L-1
Structure Extraction from Silhouette Images by Weighted Minimum Common Supergraph
Tomo Miyazaki, Shinichiro Omachi, Hirotomo Aso

-- Abstract --

It is desired to recognize objects in images which are distored or rotated. Since it is difficult to recognize those objects by template matching, recognition using structural feature has been studied. A single structure extracted by medial axis transform from a digital silhouette image is not always an essential feature that represents silhouettes in a category, because of noise and distortion. We propose a structual feature obtained by weighted minimum common supergraph as the essential feature. To show the validly of the feature, experiments are carried out for categorizing silhouette images using the features.

F3L-2
Silhouette Images Recognition by Graph Embedding in Vector Space
MingMing Zhang, Shinichiro Omachi, Hirotomo Aso

-- Abstract --

Graph representations are usually used to classify silhouette images based on topological features. A typical graph-matching algorithm is the graph edit-distance algorithm that calculates the distance as dissimilarity between two graphs. In general, multidimensional scaling method can embed objects into a vector space based on their distance information. In this paper, we present a fast method to embed graph into vector space for graph classification by using multidimensional scaling based on graph edit-distance information.

F3L-3
Image Noise Removal Based on Level Set Methods
XiaoYong Zhang, Masahide Abe, Masayuki Kawamata

-- Abstract --

This paper introduces a method for noise removal using level set methods (LSM). The method introduced in this paper relies on two fundamental ideas. Firstly, a level set formulation is used for evolving curve. Secondly, the particular form of the curvature flow is governed by a max/min switch function. Experimental results show the efficiency of LSM in the noise removal from binary image corrupted with different amounts of salt-and-pepper noise.

F3L-4
Blotch Detection Using a Background Subtraction Method for Degraded Film Sequences
Koki Shoji, Masahide Abe, Masayuki Kawamata

-- Abstract --

In this paper, a blotch detection method for degraded film sequences is proposed. The proposed method uses a background image as a reference image for blotch detection to obtain high performance when the appearance probability of blotch is high. The background image is estimated by applying filters selectively which are a double window modified trimmed mean (DW-MTM) filter, a median filter and a mean filter to the sequences in temporal direction at each pixel. The experimental result shows the proposed method is a robust blotch detection method comparing conventional methods for the high appearance probability of blotch and it is confirmed the proposed method can detect a region that blotch appeared at the same spatially location in the consecutive frame.

Go back to the Top.

F4J. Magnetic Materials and Storage Devices

Location: 3F Room J
Date & Time: 14:15-15:45, Thursday, 16 October 2008
Chair: Tomoyuki Ogawa, Assistant Professor

F4J-1
Curie temperature modulation in ferromagnetic semiconductor (Ga,Mn)As
Yu Nishitani, Daichi Chiba, Masaki Endo, Fumihiro Matsukura, Hideo Ohno

-- Abstract --

 Ferromagnetic semiconductor (Ga,Mn)As shows hole-induced ferromagnetism whose Curie temperature T_C is a function of hole concentration p. In this work, we examined that relationship between T_C and p by changing p in (Ga,Mn)As channel with field-effect transistor structure.
 Field-effect transistors were fabricated from nanometer-thick (Ga,Mn)As layer grown by molecular beam epitaxy. p was evaluated from the change of its conductance by gate electric field. Magnetic property was probed thorough the anomalous Hall resistance, and T_C was determined from the Arrott plots. We found that T_C can be scaled as T_C ∝ p^0.2.

F4J-2
The electric-field assist annealing on ferromagnetic semiconductor (Ga,Mn)As with FET structures
Masaki Endo, Daichi Chiba, Fumihiro Matsukura, Hideo Ohno

-- Abstract --

The Curie temperature T_C of a hole-inducedferromagnetic semiconductor (Ga,Mn)As can be increased bypost-annealing at around 200 ^oC, where the diffusion ofinterstitial Mn (Mn_I, ionized donor) towards the surfaceplays a role. In this work, we have examined the effect ofelectric-fields E_a on Mn_I diffusion. Wefabricated field-effect transistors with a (Ga,Mn)As channel to applyE_a and recorded the change of the sheet resistanceRsheet to monitor the Mn_I diffusion process duringannealing. The negative E_a resulted in a steeperdecrease of R_sheet than positiveE_a, which is expected to provide efficient way toraise T_C.

F4J-3
Magnetic properties of Co-Pt/Co Hard/Soft-stacked dot arrays
Kaname Mitsuzuka, Takehito Shimatsu, Nobuaki Kikuchi, Osamu Kitakami, Hiroaki Muraoka, Hajime Aoi

-- Abstract --

The magnetic properties of Co₈₀Pt₂₀(10nm)/Co(0,3, 5 nm) hard/soft stacked dot arrays with dot diameters,D, ofabout 50 nm were studied. Co-Pt films were deposited on Ru(001) seedlayers, resulting in a large uniaxial magneticanisotropy,K_u, with the c-axis perpendicular to thefilm plane.MFM images revealed that all dot arrays showed a singledomain state.The remanence coercivity, H_r, decreasedfrom 9.1 kOeto 6.9 kOe as the Co layer thickness increased from 0 to 5nm,indicating that the hard/soft stacked structure was effectiveatreducing the switching field, as theoretically predicted.

F4J-4
Reinforcement of Exchange Biasing of Ultra Thin Antiferromagnetic Oxide Layer
Kazuya Sawada, Masaaki Doi, Masashi Sahashi

-- Abstract --

Magnetic oxide layer has attracted attention for various application fields, e.g., exchange biasing from antiferromagnetic oxide with magnetoelectric effect which has studied for controlling ferromagnetic spins. However, the blocking temperature (T_B) decreases with thinning the thickness. We have obtained higher T_B of 140K for CoO-FeO NOL (Nano-Oxide-Layer) and 220K for Cr₂O₃-Fe₂O₃ NOL in spite of those thicknesses are only 1nm. Why such ultra thin oxide shows a high T_B and a valuable exchange bias is interesting issue. Therefore, in this study the temperature dependencies of the exchange bias field for CoFe/NOL/CoFe trilayer and IrMn/CoFe/NOL/CoFe SV are investigated for discussion.

F4J-5
Enhancement of magnet properties of granular perpendicular recording media by using a pseudo-hcp nonmagnetic intermediate layer
Shintaro Hinata, Shin Saito, Migaku Takahashi

-- Abstract --

Materials with the fcc structure with stacking faults (pseudo-hcp) are examined for the nonmagnetic intermediate layer (NMIL) in order to have high perpendicular magnetic anisotropy energy (K_u) on granular-type perpendicular recording media. In this article, K_u of the overlying medium is found to be enhanced by loss of stacking order on the (111) plane in fcc NMIL. One persuasive reason for this enhancement is the preferred epitaxial growth of magnetic grains on only the (111) atomic terrace of the fcc NMIL caused by collapse of the sixfold symmetry of (-111), (1-11), and (11-1) atomic terraces of the fcc NMIL.

F4J-6
Observation of Fullerene Molecules on Si(111)-7x7 Reconstructed Surface by Non-contact Scanning Non-linear Dielectric Microscopy
Shin-ichiro Kobayashi, Yasuo Cho

-- Abstract --

Fullerene (C60) molecules deposited on the Si(111)-7×7 (7×7) reconstructed surface have been investigated by the non-contact scanning non-linear dielectric microscopy (NC-SNDM) under ultra-high vacuum condition. Both topography and the internal structure of the C60 molecule on the 7×7 surface have been successfully resolved. Some hexagonal and pentagonal rings of the C60 molecule on the three-hold symmetric site of the on the 7×7 surface are slightly resolved. This suggests that the NC-SNDM technique can resolve the atomic of the C60 molecules.

Go back to the Top.

F4K. Semiconductor Devices

Location: 3F Room K
Date & Time: 14:15-15:45, Thursday, 16 October 2008
Chair: Yuichiro Tanushi, Assistant Professor

F4K-1
Strain Control of Si/Si_1-x-yGe_xC_y/Si(100) Heterostructure by Stripe-Shape Patterning
Kiyeol Kim, Masao Sakuraba, Junichi Murota

-- Abstract --

Strain control of Si/Si_1-x-yGe_xC_y/Si(100) heterostructure by stripe-shape patterning was investigated. Stripe width dependence of Raman shift for Si-Si vibration mode in the deep-submicron-width stripes becomes obviously weaker at the higher C fraction. Moreover, position dependence of Raman shift measured for a 4.6 um-width stripe indicates that the change in Raman shift and the width of a strain relaxed region near the edge are smaller at the higher C fraction. These results suggest that the effective propagation length of strain relaxation at the same thickness becomes smaller by C introduction.

F4K-2
P Atomic-Layer Doping in Heteroepitaxial Growth of Si/Si_1-xGe_x(100) by Ultraclean Low-Pressure CVD
Yohei Chiba, Masao Sakuraba, Junichi Murota

-- Abstract --

Effects of Si epitaxial growth condition upon P doping characteristics in Si₂H₆ reaction on P atomic-layer adsorbed Si_0.3Ge_0.7(100) using ultraclean low-pressure CVD have been investigated in order to increase carrier concentration in group IV semiconductor. It is found that, with increasing Si₂H₆ partial pressure, the Si growth rate at 450 ^oC tends to increase and surface P segregation tends to be reduced. The result is important to realize extremely high carrier concentration locally in the nanometer-order heterostructure.

F4K-3
Research on Poly-Si TFTs with One-Dimensionally Long Grains Formed by CW Laser Lateral Crystallization
Xiaoli Zhu, Shin-Ichiro Kuroki, Shuntaro Fujii, Masayuki Numata, Koji Kotani, Takashi Ito

-- Abstract --

Carrier transport mechanism in polycrystalline silicon (poly-Si) thin film transistors (TFTs) with one-dimensionally long grains of 20x2 um2in size formed by continuous wave (CW) laser lateral crystallization was investigated. Effective carrier mobility in the poly-Si TFTs was mainly governed by phonon scattering as in bulk Si transistors. It was also found that surface orientation fluctuation of each grain caused the mobility fluctuation ruled by central limit theorem.

F4K-4
Three-Dimensional (3D) Integration of A Log Spiral Antenna for High-Directivity Plasmon-Resonant Terahertz Emitter
Hyun-Chul Kang, Takuya Nishimura, Taiichi Otsuji, Naoya Watanabe, Tanemasa Asano

-- Abstract --

We have designed and fabricated a high-directivity plasmon-resonant terahertz emitter (PRE) having a 3D-integrated antenna complex. Previous PRE is based on a high electron mobility transistor (HEMT) and incorporates doubly interdigitated grating gates of broadband terahertz antennae converting non-radiative plasmons to radiative electromagnetic waves. Due to sub-wavelength aperture of practical grating dimensions, however, undesirable diffraction deteriorates the radiation directivity. A fine-patterned planar log-spiral antenna is dedicated on a separate Si substrate which is tightly coupled with the PRE in a 3D stack via micro-bump connections. The FDTD analysis demonstrates a better directivity by ~70% than that for conventional in the terahertz range.

F4K-5
The Bias Dependence on 2-Dimensional Plasmon-Resonant Terahertz Emitter Based on a Novel Super-Grating Dual-Gate Structure
Takuya Nishimura, Nobuhiro Magome, HyunChul Kang, Taiichi Otsuji

-- Abstract --

Recently, we’ve proposed a 2-dimensional plasmon (2DP) resonant emitter (PRE). Since the drain bias causes a linear potential slope from the source to drain, the electron densities in 2DP cavities are dispersed. As a result, the 2DPRE exhibits undesirable spectral broadening. A novel super-grating gate can compensate for the electron density distribution by modulating the grating dimension. The finite difference time domain (FDTD) simulation confirms spectral narrowing by optimizing the dimension of the super-grating gate. When the gate and drain bias are detuned from an optimum condition, the super-grating gate can maintain a certain amount of the spectral narrowing effect.

F4K-6
Study of Self-Heating Phenomena in Si Nano Wire MOS Transistor
Kousuke Tanaka, Tetsuo Endoh

-- Abstract --

In this study, I have numerically investigated thetemperaturedistribution of n-type Si Nano Wire MOS transistor (NW-MOSTr.)induced by the self-heating effect by using a 3-D devicesimulator.The dependencies of temperature distribution within theNW-MOS Tr. onboth its gate length and width of the Si nano wire wereanalyzed. Itis found that the increase of its peak temperature due toscaling thegate length can be suppressed by scaling the size of the Sinano wire,for the first time. This study shows very useful results forfutureNW-MOS Tr. design for suppressing the self-heating effect.

Go back to the Top.

F4L. Image Processing II and Display Device

Location: 3F Room L
Date & Time: 14:15-15:15, Thursday, 16 October 2008
Chair: Masahide Abe, Associate Professor

F4L-1
A Robust 3D Face Recognition Algorithm Using Passive Stereo Vision
Akihiro Hayasaka, Koichi Ito, Takafumi Aoki, Hiroshi Nakajima, Koji Kobayashi

-- Abstract --

The recognition performance of the conventional 3D face recognition algorithm using ICP (Iterative Closest Point) is degraded for the 3D face data with expression changes. Addressing this problem, we consider the use of the expression-invariant local regions of a face. We find the expression-invariant regions through the distance analysis between 3D face data with the neutral expression and smile, and propose a robust 3D face recognition algorithm using passive stereo vision. We demonstrate efficient recognition performance of the proposed algorithm compared with the conventional ICP-based algorithm through the experiment using a stereo face image database which includes the face images with expression changes.

F4L-2
An Image Correction Method for Multi-Projector Display Using Image Feature Points
Norihito Numa, Toru Takahashi, Tatsuya Kawano, Takafumi Aoki, Satoshi Kondo

-- Abstract --

This paper proposes an image correction method for multi-projectordisplay using SIFT image feature points. A projected image correctionmethod based on projector-camera system needs to obtain thecorrespondence between projector and camera images. The accuratecorrespondence is required to achieve seamless imagery with multipleprojectors. In conventional methods, the correspondence is obtained byprojecting the special patterns on a screen. On the other hand, weobtain the correspondence using SIFT feature matching. Throughanexperiment, we demonstrate that the proposed method is effective forthe practical use of projector-camera systems compared with theconventional methods.

F4L-3
A Geometric Correction Method Using Stereo Camera for Projected Images
Toru Takahashi, Norihito Numa, Takafumi Aoki, Satoshi Kondo

-- Abstract --

This paper proposes a geometric correction method using stereo camera for projected images. When we use a projector in the various places in the real world, a projected image is distorted due to the shape of a display surface, the positional relationship between a projector and a display surface, etc. Addressing this problem, we propose a geometric correction method for projected images, which employs 3D measurement technique with the narrow-baseline stereo camera. In the proposed method, we use the correspondence search technique based on Phase-Only Correlation (POC) to obtain the high-accuracy 3D shape of the display surface. Experimental results show the proposed method is effective for correcting distortion of projected images.

F4L-4
Analysis of Speckle on the screen of laser light projection display
Yuhei Kuratomi, Baku Katagiri, Tohru Kawakami, Tatsuo Uchida

-- Abstract --

Recently laser projection displays are drawing considerable attentionsdue to high image qualities and low energy consumption. When usinglasers for projections, a high contrast, fine-scale granular patternis seen by an observer looking at the screen. The pattern is calledspeckle and has been a major problem for the application of laserprojections. Reduction methods of speckles applicable to actualproducts are strongly demanded. As an important step to the solution,we deduced a model of the occurrence of speckles in terms of displayapplication, and analyzed the effects of the speckle reduction of themoving diffuser method.

Go back to the Top.

F5P. Poster Session

Location: 3F Room P
Date & Time: 16:00-17:30, Thursday, 16 October 2008
Chair: Shunsuke Koshita, Assistant Professor

F5P-1
Complex BPF for Low-IF Receivers with Automatic Digital Tuning Circuit
Jun Gyu Lee

F5P-2
Study of Broadband MIMO Communication System and Multi-antenna
Takanori Iwaki, Jing Xin Fu, Yohei Kojima, Takahiro Chiba, Takafumi Fujimori, Yu Kakinuma, Kazuki Saito, Lin Wang

F5P-3
A Study on the New End-to-End Congestion Control for Satellite IP Networks
Satoshi Utsumi, George Serban Radescu

F5P-4
Research for Achievement of High-Quality Displays
Yuji Ohno, Yuhei Kuratomi, Se-Joon Oh

F5P-5
A Study on Effective Application-Layer Multicast toward Intelligent Information and Communication Services
Toshiaki Osada, Jun Makishi

F5P-6
Open 3D shared space construction framework for next generation collaborative work
Takayuki Kuroda, Akira Sakatoku

F5P-7
Design of linear digital filters for realization in hardware
Ken Kouno

F5P-8
Low-power hydrogen gas sensor using 3C-SiC/Si hetero-structure
Eiji Saito, Roland Bantaculo

F5P-9
Research on network management framework for intelligent information and communication services
Akihiro Satoh, Masahiro Nagao

F5P-10
Formation of Ferromagnetic Nanocontacts in Cr-NOL
Kazuya Sawada, Shohei Kawasaki, Masaaki Doi, Masashi Sahashi

F5P-11
Synthesis and properties of nitrogen included fullerenes using electron-temperature-gradient controlled plasma
Jamal Uddin Ahamed, M M Rhaman, M Miyanaga, S Tamura

F5P-12
A study on development of a high accurate automatic digital restoration system in old film sequences
Koki Shoji, Xiao Yong Zhang, Sang-Churl Nam

F5P-13
Synthesis of carbon nanotubes encapsulating magnetic metal using hybrid advanced plasma
Zohreh Ghorannevis, Kazuhiko Baba

F5P-14
Creation of High-Carrier-Concentration and High-Mobility New Group IV Semiconductors for Future Device
Yohei Chiba, Kiyoel Kim, Tomoyuki Kawashima, Takayuki Nosaka, Tomoya Hirano, Atsushi Uto, Kuniaki Takahashi, Masao Sakuraba, Junichi Murota

F5P-15
Optical-fiber cross-Kerr nonlinearity with less than one photon
Nobuyuki Matsuda, Jun Zhang, Wakana Ueno, Ryosuke Shimizu, Yasuyoshi Mitsumori, Hideo Kosaka, Keiichi Edamatsu

F5P-16
Research on Terahertz Emitter Utilizing Plasmonic Dispersion Management
Takuya Nishimura, Kang HyunChul

F5P-17
Fabrication of Co-Pt perpendicular magnetization film with a magnetic anisotropy suitable for STT microwave oscillation element
SeungMo Noh

F5P-18
A Study on Real-World Signal Processing and Its Applications
Akihiro Hayasaka, Norihito Numa, Atsushi Miyamoto, Toru Takahashi

F5P-19
Formation of new nanomaterials on Si substrates applicable to Si technology
Hiroyuki Handa, Yu Miyamoto

F5P-20
Optical Switches and Optical Buffer Designs for High-Speed Optical Networks
Yusuke Fukushima, Xiaoliang Wang, Dalia Mohamed Nashat, Xiaohong Jiang, Susumu Horiguchi

F5P-21
Research on High Performance Poly-Si TFTs with One-Dimensionally Long Grain and Reduced Surface Roughness
Shuntaro Fujii, Masayuki Numata, Koji Hara, Masaki Midorikawa, Shin-Ichiro Kuroki, Koji Kotani, Takashi Ito

F5P-22
Device, Circuit, and Architecture Technologies toward a High-performance VLSI System
Shoun Matsunaga, Daisuke Suzuki, Hirokatsu Shirahama

F5P-23
Fabrication of thin film transistors (TFT) for flexible display
Mitsutaka Matsumoto, Yohei Inayoshi, Syougo Murashige, Yoshihisa Yamamoto, Noriyuki Toyoda, Yasushi Suzuki

F5P-24
Control of spin-related properties in semiconductors by modulating career density
Tsung-Tse Lin, Yu Nishitani, Masaki Endo, Masaaki Ono

F5P-25
Sabotage-Tolerance Mechanisms for Reliable Volunteer Computing Systems
Kan Watanabe, Masaru Fukushi, Susumu Horiguchi

F5P-26
Specialty optical waveguides and fibers for medical and biosensing applications
Tomonori Watanabe, Lin Ma, Yoshio Morales

F5P-27
AC loss investigation of a tri-axial HTS cable with balanced current distribution
Ahmet Nuri Ozcivan, Seiji Soeda, Hu Nannan, Tsuyoshi Yagai, Makoto Tsuda, Takataro Hamajima

F5P-28
Clarification of the auditory-visual information processing mechanism for advancing communication systems
Dae-Gee Kang, Kaori Asakawa, Royta Miyauchi, Shuichi Sakamoto, Yukio Iwaya, Yôiti Suzuki

F5P-29
Theoretical considerations of determining accurate acoustical physical constants of Ca₃NbGa₃Si₂O₁₄ and Ca₃TaGa₃Si₂O₁₄ single crystals
Takashi Shishitani, Yusuke Kourai, Shoy Yoshida, Mototaka Arakawa, Yuji Ohashi, Jun-ichi Kushibiki

F5P-30
Influence of Strain Effects on Characteristics of Nano-grating channel MOSFETs
Xiaoli Zhu, Shin-Ichiro Kuroki, Jun Jiang, Toshiyuki Abe, Yuji Ishinaga, Koji Kotani, Takashi Ito

F5P-31
Delay-Minimized Load Distribution for Multipath Networks
Sumet Prabhavat, Hiroki Nishiyama, Nei Kato

F5P-32
High-Performance Wireless Ad Hoc Networks
Kaikai Chi, Xiaohong Jiang, Susumu Horiguchi

F5P-33
New approaches to the network-oriented information technlogies for the pleasant sound communications
Seongjun Hahm, Takuma Okamoto, Hoseok Wey, Yôiti Suzuki, Shozo Makino

F5P-34
A Secure and Robust Framework of Service Level Specification for Next Generation Wireless Applications
Zubair Md. Fadlullah, Juan Carlos Fernandez, Nei Kato

F5P-35
Investigation of the Design Principles for Adaptive Intelligence
Wataru Watanabe, Takuya Umedachi, Dai Owaki

Go back to the Top.

S1J. Invited Lecture V

Location: 3F Room J
Date & Time: 09:00-10:00, Friday, 17 October 2008
Chair: Yuzo Ohno, Associate Professor

Spin Dynamics in Ferromagnetic and Spintronic Materials
Can-Ming Hu
Department of Physics and Astronomy, University of Manitoba, Canada

-- Abstract --

Understanding spin dynamics in ferromagnetic and spintronic materials has both scientific significance and application importance [1]. It forms an exciting frontier of condensed matter physics that provides excellent research opportunities for motivated graduate students. Within this context, I will review in this talk our recent work performed at University of Manitoba, Canada, which include: generation of electrical current via spin precession [2], electrical detection of spin waves in permalloy microstrips [3], direct mapping of spin wave evolution, broadband ferromagnetic resonance spectroscopy on GaMnAs [4], and developing a new approach for detecting RF magnetic field vectors [5].

Go back to the Top.

S1K. Invited Lecture VI

Location: 3F Room K
Date & Time: 09:00-10:00, Friday, 17 October 2008
Chair: Xiaohong Jiang, Associate Professor

Spectrum Sensing, Sharing and Management in Cognitive Radio
Jiandong Li, Dong Chen
State Key Laboratory of Integrated Service Networks, Xidian University, China

-- Abstract --

Cognitive radio is a potential enabling technique for high efficient frequency utilization and End-to-End efficiency ofwireless networks. In this article, we give a briefoverview of cognitive radio in the context of opportunisticspectrum access and its key technologies and challenges. Thedevelopments in spectrum sensing, sharing and managementare discussed in detail.

Go back to the Top.

S1L. Invited Lecture VII

Location: 3F Room L
Date & Time: 09:00-10:00, Friday, 17 October 2008
Chair: Masanori Hariyama, Associate Professor

Technical Mechanisms to Protect Chip Design Intellectual Property
Tom Kean
Algotronix Ltd.

-- Abstract --

Intellectual Property theft in the semiconductor industry takes many forms: counterfeit chips, illegally copied FPGA bitstreams and illegal use of IP cores or CAD software. The problem has reached epidemic proportion over recent years as the electronics industry has focussed on cost reduction and manufacturing has been outsourced to countries with weak intellectual property enforcement. This talk will consider various technical mechanisms to protect intellectual property, including DesignTag, a small, low power IP core which allows intellectual property within an operating integrated circuit to be reliably identified simply by placing a sensor in contact with the chip package.

Go back to the Top.

S2J. Spintronics and Quantum Electronics

Location: 3F Room J
Date & Time: 10:15-11:15, Friday, 17 October 2008
Chair: Yuzo Ohno, Associate Professor

S2J-1
Gate voltage dependence of nuclear spin relaxation time in GaAs quantum well
Masaaki Ono, Shunichiro Matsuzaka, Yuzo Ohno, Hideo Ohno

-- Abstract --

Manipulation and detection of nuclear spins insemiconductornanostructures have been of great interest not only forunderstandingthe physics but also to implement quantum informationdevices. In thiswork, we investigated the electron-density dependenceof the nuclearspin relaxation and decoherence times in aSchottky-gatedn-GaAs/AlGaAs quantum well (QW) by time-resolvedKerr-rotationmeasurements combined with pulsed-rf nuclear magneticresonance. Weshow that the nuclear spin relaxation and decoherencetimes decreasewith decreasing electron density, indicating that thehyperfineinteraction is enhanced as the electronic states becomeslocalized inan impurity-doped QW.

S2J-2
Temperature dependence of threshold current density of 3.8 THz GaAs-based quantum cascade lasers
Tsung-Tse Lin, Keita Ohtani, Hideo Ohno

-- Abstract --

Quantum cascade lasers (QCLs) utilizing carrier recycling are expected to be one of the good candidates for THz sources because they provide high optical output power. But the maximum operation temperature is 178 K, which limits their applications such as spectroscopy and imaging. One of the alternatives to improve operation temperature is to increase the characteristic temperature of threshold current density (Jth). Here we fabricate 3.8 THz GaAs QCLs and study the temperature dependence of Jth. By comparing the calculation based on a linear rate equation, temperature degradation mechanism of Jth is discussed.

S2J-3
Spin Transfer Torque Assited Magnetization Switching in Perpendicular Magnetized Thin Film
SeungMo Noh, Masaaki Doi, Masashi Sahashi

-- Abstract --

In this paper, we present a fabrication of Co-Pt perpendicular magnetization film suitable for STT microwave oscillation element. The device consist of four spin valve elements, each of which comprises of a transparent top electrode, free and pinned magnetic layers separated by a copper space, perpendicular magnetized layer and bottom electrode. Co-Pt ordered alloy films with a magnetic anisotropy, Ku, were successfully fabricated at relatively low substrate temperatures using ultrahigh vacuum sputter film deposition. The ordered Co-Pt films, with the <111> direction (easy axis of magnetization) perpendicular to the film plane, were fabricated on Si substrates. Despite a relatively large device resistance, spin transfer switching of the free layer will be carried out using the spin injection from the transparent top electrode.

S2J-4
Electrical field effect of trions in GaAs quantum well structures
Wakana Ueno, Hideo Kosaka, Takeshi Kutsuwa, Makoto Kuwahara, Yasuyoshi Mitsumori, Keiichi Edamatsu

-- Abstract --

We studied photoluminescence properties of charged excitons in a GaAs quantum well in order to detect the quantum state of a single electron spin using an optical method. For this purpose, we used an opaque Ti mask evapolated as a Schottky electrode with optical apertures. When biased with negative voltages, the gates create potential for electrons. We estimate potential energy in real space which depends on the size of optical apertures as well as the bias voltage.

Go back to the Top.

S2K. Signal Processing

Location: 3F Room K
Date & Time: 10:15-11:45, Friday, 17 October 2008
Chair: Shinichiro Omachi, Associate Professor

S2K-1
Experiment and Evaluation of Iterative Construction Method of Application-Layer Multicast Tree for Many-to-Many Communication
Toshiaki Osada, Jun Makishi, Gen Kitagata, Tetsuo Kinoshita, Norio Shiratori

-- Abstract --

We have proposed an iterative construction method ofApplication-LayerMulticast (ALM) tree, which improves unfairness ofcommunicationquality in a stepwise manner among end-nodes in ALM. Inthis paper, webriefly introduce the proposed method and then evaluatetheperformance of it through the computer simulation. Thesimulationresults show that the proposed method can increase theminimumreceiving bit-rate in total of the end-nodes up to 1.31 times.It canalso reduce standard deviation of all by one-quarter, comparedto theexisting method.

S2K-2
TCP-Cherry: A New Scheme of TCP Congestion Control for Satellite IP Networks
Satoshi Utsumi, Salahuddin Muhammad Salim Zabir, Norio Shiratori

-- Abstract --

In this paper, we introduce a new TCP congestion control mechanism for satellite IP networks, TCP-Cherry. TCP-Cherry uses low-priority data segments, namely, supplement segments that prove the network available resouces as wellas carry new data blocks not yet transmitted as regular data segments. Our new congestion control algorithms, Fast-Forward Start and First-Aid Recovery use the supplement segments. A unique charactaristics of our scheme is the mechanism for supplementsegment keeping the overhead of duplicate transmissions minimum. Simulation results show that TCP-Cherry yields upto a maximum improvement of more than 150% in goodput compared with other existing TCP congestion control schemes.

S2K-3
A novel method to grasp the detected events in event-based management
Akihiro Satoh, Masahiro Nagao, Kazuhide Koide, Norio Shiratori, Tetsuo Kinoshita

-- Abstract --

The analysis of events is indispensable in event-based management for TCP/IP network. However, when large amount of events are generated, overlooking of minor events or repetition of analyzing the same events could be a serious concern. In this paper, we propose a novel method to evaluate the degree of importance of an event. The degree is based on the amount of information that can be obtained by analyzing the traffic data of the event generated. Thus, an efficient way to detect the insights of events is possible and as a result improvement in network management can be achieved.

S2K-4
Asynchronous data-transfer interface for an interleaver in fully-parallel low-density parity-check decoders
Naoya Onizawa, Takahiro Hanyu

-- Abstract --

We present an asynchronous data-transfer interface for an interleaverin fully-parallel low-density parity-check decoders. Using theasynchronous interleaver realized by multiple-valued encoding andcircuits, each message's arrival rate between computation nodes isbased on the delay due to the wire length without overhead of wiring.Therefore, the throughput is not restricted by the worst-case latencyin the synchronous interleaver, which results in a higher average rateof computation. Moreover, since the mode-detection circuit thatobserves data-transmission modes; full-duplex, half-duplex and standbymodes, is added to the data-transfer circuit, current sources can becorrectly cut off, which reduces the power dissipation.

S2K-5
Design of FIR digital filters with monotonic transition bands
Ken Kouno, Masahide Abe, Masayuki Kawamata

-- Abstract --

In this paper we propose an iterative technique forthefrequency-domaindesign of finite-impulse-response (FIR)digitalfilters with monotonictransition-band magnituderesponses.Transition-band monotonicity isguaranteed by imposingconstraints onthe first-derivatives of amplituderesponses.FIR filterdesign subject to the first-derivative constraintsisformulated as aniterative polynomial approximation problem withjusta few linearconstraints.FIR filters with monotonic transitionbands can then beobtained byapplying the proposed technique toconventionalapproximation methods.An example design is included toillustrate theeffectiveness of theproposed technique.

S2K-6
Synthesis of Minimum L₂-sensitivity Realization without Limit Cycles in State-Space Digital Filters
Shunsuke Yamaki, Masahide Abe, Masayuki Kawamata

-- Abstract --

This paper proposes a synthesis method of limit cycle free state-spacedigital filters with minimum L₂-sensitivity. Thecoordinate transformation matrix which transforms the balancedrealization into the minimum L₂-sensitivityrealization is expressed as the product of a positive definitesymmetric matrix and an arbitrary orthogonal matrix. We prove that thecontrollability and observability Gramians of the minimumL₂-sensitivity realization satisfy a sufficientcondition for the absence of limit cycles if we select an appropriateorthogonal matrix. As a result, the minimumL₂-sensitivity realization without limit cycles canbe synthesized.

Go back to the Top.

S2L. Multimedia Information Processing

Location: 3F Room L
Date & Time: 10:15-11:45, Friday, 17 October 2008
Chair: Toru Abe, Associate Professor

S2L-1
Packet loss concealment of audio signal using multiple description audio coding based on time domain division.
Hoseok Wey, Akinori Ito, Yôiti Suzuki

-- Abstract --

Multiple description (MD) coding separates the source signal intomultiple sub-data, called descriptions. A signal perceived similar tothe source signal can be generated using one of the multiple descriptions. The MD coding is useful for packet loss concealment ofdigital media over IP network. In this paper, we describe an MD codingmethod of audio signal, where the source signal is split into odd andeven samples and compressed by MP3 encoders separately. The generateddescriptions are decoded by MP3 decoders and compensated by a Wienerfilter. We conducted experiments to compare the proposed method toseveral conventional methods.

S2L-2
Hybrid auralization of concert-hall acoustics using scale model measurement and geometric acoustic simulation
Makoto Otani, Yukio Iwaya, Fumitaka Saito, Yôiti Suzuki, Yoshinari Yamada, Takayuki Hidaka

-- Abstract --

The authors propose a novel hybrid auralization method forexperiencing acoustic characteristics of large spaces such as concerthalls. This method consists of scale-model measurement and computersimulation of acoustic field. The measurement employs a scale model of a hall and ahigh-precision scale-dummy-head microphones which predicthall-acoustics accurately at lower frequencies, whereas the simulationbased on geometrical acoustic simulation predicts hall-acousticsaccurately at higher frequencies. In other words, the hybridauralization method takes advantage of a precisely predicted part ofeach method. This paper introduces the notion of this new auralizationmethod, procedures of measurement, simulation method, and the resultsof the proposed method.

S2L-3
Temporal recalibration in audio-visual speech integration using temporal order judgment
Kaori Asakawa, Akihiro Tanaka, Hisato Imai, Shuichi Sakamoto, Yôiti Suzuki

-- Abstract --

Audio-visual synchrony is very important for comfortable speechcommunications. Previous studies have revealed a temporal windowduring which human observers perceive physically desynchronizedauditory and visual signals as synchronous in both speech andnonspeech signal. This temporal window of audio-visual integration isrecalibrated after adaptation to a constant timing difference betweenauditory and visual signals in nonspeech. In this study, weinvestigate whether or not the after-effects of the consequences oftemporal recalibration occurs even in speech stimuli. Our resultssuggest that the temporal recalibration occurs not only in nonspeechsignal but also in monosyllabic speech.

S2L-4
Open 3D Shared Space Combination Scheme for Symbiotic Collaborative Works
Takayuki Kuroda, Akira Sakatoku, Gen Kitagata, Norio Shiratori

-- Abstract --

In this paper, we propose an open 3D shared space combination scheme based on the concept of Symbiotic Computing. Existing 3D collaborative work supporting systems have some limitations with their 3D shared space construction because these systems have basically closed nature. To break these limitations, we are developing symbiotic collaborative work supporting system which can effectively support remote collaborative work. Our scheme enables us to easily and flexibly build collaborative work environment by reusing 3D structure definitions designed by the general public. In addition, we confirm effectiveness of our proposed scheme through experiments by using a prototype system.

S2L-5
Open Application Integration Scheme based on Distributed Partial Space Combination
Akira Sakatoku, Takayuki Kuroda, Gen Kitagata, Norio Shiratori, Tetsuo Kinoshita

-- Abstract --

In this paper, we propose open application integration scheme based ondistributed partial space combination. Existing systems havelimitation with reusability of the application's function, becauseexisting application integration schemes are closed. To break thislimitation, we developing open application integration scheme bytransmission of application's function itself. Our scheme enabledusers to reuse and integrate applications designed by a number ofdesigners. In addition, we show design and implementation of theprototype system, and confirm the effectiveness of our scheme byseveral experiments.

S2L-6
An evaluation method depending on the feature of comment in BBS
Yu Ichifuji, Susumu Konno, Hideaki Sone

-- Abstract --

Electronic bulletin board system (BBS) has a problem of abuse such as ruined debate. The authors had proposed a method that aids an operator to detect such abuse.The method employs ``Ruination Figure [RF]'' which indicates influence of a comment in the BBS,and this method was able to detected 81% of such abuse.To improve the method a problem, the authors realize that problem concerning a vague criteria is attributable to that each influence of words is treated as a constant in this method.The authors discuss the validity of weight of words.

Go back to the Top.

S3J. Digital Processing and Biological Devices

Location: 3F Room J
Date & Time: 13:00-14:15, Friday, 17 October 2008
Chair: Hiroyuki Nakase, Associate Professor

S3J-1
Complex BPF for Low-IF Receivers with Automatic Digital Tuning Circuit
Jun Gyu Lee, Shoichi Masui, Hideaki Kondo, Masaru Sawada, Norio Murakami

-- Abstract --

An implementation of automatic digital tuning circuit for complex BPF in low-power and low-cost transceivers are explored. The principle of tuning through the calibration of a capacitor array is illustrated. The developed tuning technique achieves a tuning error below ±3.5%.

S3J-2
Spurious Control Method for FBAR
Shoichi Tanifuji, Yuji Aota, Hiroshi Oguma, Suguru Kameda, Tadashi Takagi, Kazuo Tsubouchi

-- Abstract --

The band pass filters based on the film bulk acoustic resonators (FBAR) have a great advantage of the good cut off characteristics at over 2GHz. We have fabricated 5GHz FBAR using aluminium nitride (AlN) by the metal organic chemical vapor deposition (MOCVD) method. However, there was the case that a large spurious existed between the resonant frequency and the anti-resonant frequency on the admittance characteristics of the prototype FBAR. When we consist the band pass filters, this spurious will become major problems. Therefore, we analyzed the spurious control method using the three-dimensional (3−D) finite element method (FEM) simulation.

S3J-3
AlN Film Deposition using MOCVD Method for FBAR
Yuji Aota, Shoichi Tanifuji, Hiroshi Oguma, Suguru Kameda, Tadashi Takagi, Kazuo Tsubouchi

-- Abstract --

We have investigated a 5 GHz band film bulk acoustic resonator (FBAR) for wireless local area network (WLAN) filters. The FBAR characteristics mainly depend on the piezoelectric film crystallization. An AlN was used as a piezoelectric material because of its high phase velocity, good temperature coefficient of delay and moderate piezoelectric coupling coefficient. We used a metal organic chemical vapor deposition (MOCVD) method and achieved high oriented AlN(0002) film at the FBAR structure.

S3J-4
High efficiency dental excavation by dual-wavelength laser radiation
Tomonori Watanabe, Yuji Matsuura

-- Abstract --

For highly efficient excavation of dental hard tissues, we propose simultaneous radiation of Ho:YAG (λ=2.1 μm) and Er:YAG (λ=2.94 μm) lasers. Er:YAG lasers offer high ablation rate for dental hard tissues because of the laser wavelength that coincides with a absorption peak of water. In contrast, Ho:YAG lasers show reasonably high ablation rate with low induced heat. When radiating these two lasers simultaneously or with a slight interval, dental hard tissues can be excavated with high ablation rate and low induced heat.

Go back to the Top.

S3K. Wireless Communications II and Antennas

Location: 3F Room K
Date & Time: 13:00-14:15, Friday, 17 October 2008
Chair: Suguru Kameda, Assistant Professor

S3K-1
Measurement of Indoor Multipath Propagation Characteristics at 60 GHz with High Delay Time Resolution
Hao Ngoc Tran, Suguru Kameda, Tadashi Takagi, Kazuo Tsubouchi

-- Abstract --

In this work, delay profiles of indoor multipath characteristicsat 60GHz were measured by using a pulse compression method basedon M-sequences. Packets included preamble of two M-sequenceswere generated by a software modulator and a 4 GSample/s dualchannel arbitrary waveform generator. The delay time resolution was0.25 ns. Arrival angles of multipaths were measured by rotating anarrow beam received antenna with angle resolution of 5 degrees. Frommeasurement results, the multipaths were found to be separable. Themultipaths can be combined by using a rake receiver.

S3K-2
Effects of gain imbalance of array elements on MIMO channel capacity
Takanori Iwaki, Yu Kakinuma, Qiang Chen, Kunio Sawaya

-- Abstract --

Gain imbalance effects of array elements on the MIMO channel capacity have been experimentally studied. 2×2 MIMO system which have two transmitting antennas and two receiving antennas is considered. Experiment is carried out in indoor environment, and antenna gain at the receiver is artificially controlled. As a result, it is found that higher performance is obtained by using equal gain antennas than unbalanced gain antennas for the design of antennas in mobile handset.

S3K-3
A Measurement Method Using a Modulated Probe Array for Phase of Electromagnetic Field
Toru Mizukami, Qiang Chen, Kunio Sawaya

-- Abstract --

A simultaneous measurement method using a parallel modulated probe array is proposed to measure the electromagnetic radiation rapidly. The performance of the system using this method has been demonstrated by the measurement time and some experimental results such as the radiation pattern and the antenna efficiency by our group. However, the phase measurement is not possible by using this system because of its structure. In this report, a phase measurement method which is using digital signal processing is proposed and some results of verification experiments are demonstrated.

S3K-4
On the Characteristics of Radome Enclosed Archimedean Spiral Antennas
Mang He, Qiang Chen, Qiaowei Yuan, Kunio Sawaya

-- Abstract --

The effects of the superspheroidal radomes on the characteristics of Archimedean spiral antennas are investigated in detail, which may provide some useful insights in the design of the radome-antenna system in practical engineering. All interactions between the radome and antennas are rigorously included in the coupled surface integral equation (CSIE) model, which leads to more accurate results than those of the existing methods based on the high-frequency algorithm or the decoupled radome-antenna analysis procedure especially when the antennas are placed in the close proximity of the radome. Numerical results show that the performance of the spiral antenna is significantly changed due to the presence of the radome.

S3K-5
Study on Dispersion Errors of Novel Hybrid CIP/FDTD Method
Jerdvisanop Chakarothai, Qiang Chen, Kunio Sawaya

-- Abstract --

Although the contrained interpolation profile (CIP) method hasasmaller numerical dispersion when compared with the standard Yee'sfinite difference time domain (FDTD) algorithm, theelectric andmagnetic field arrangements of CIP method make it difficult to modelantennas into the analysisspace. A hybridization of CIP method andFDTD method is proposed to solve the problems of large- orlong-distancepropagation space including antennas in this paper. Theboundary conditions between CIP and FDTD method are also consideredandvalidated. The results show that the numerical dispersion of thehybrid CIP-FDTD method is almost same asthat of the CIP method, whichis superior to that of the FDTD method.

Go back to the Top.

S3L. Circuit Design and Computing

Location: 3F Room L
Date & Time: 13:00-14:15, Friday, 17 October 2008
Chair: Masanori Natsui, Assistant Professor

S3L-1
Impact of a TMR-Based Nonvolatile Logic-in-Memory Circuit
Shoun Matsunaga, Takahiro Hanyu

-- Abstract --

This paper presents a tunneling-magneto-resistance (TMR)-based logic-in-memory circuit, where nonvolatile storage functions are distributed over a logic-circuit plane, whose property is demonstrated for a standby-power-free VLSI with “quick” wake-up/sleep functions without reload/write-back from/into an off-chip nonvolatile storage system, respectively. Its basic behavior and usefulness in a VLSI system are described in comparison with the conventional approach.

S3L-2
Design of a processing element based on multiple-valued current-mode logic for a many-core processor
Hirokatsu Shirahama, Takahiro Hanyu

-- Abstract --

A new processing element (PE) using multiple-valued current-mode (MVCM) circuitry is proposed for a high-performance many-core processor. A two-bit addition which is the critical path in the PE is attributed to a four-valued one-digit addition that is directly performed by using MVCM circuitry. A two-bit data can be transmitted between processing elements by a single wire with a four-level current signal. The efficiency of the proposed circuit is demonstrated using HSPICE simulation under a 90nm CMOS technology in comparison with a corresponding CMOS implementation.

S3L-3
Design of a Bit-Serial Reconfigurable VLSI Based on Multiple-Valued Control Scheme
Nobuaki Okada, Michitaka Kameyama

-- Abstract --

A fine-grain bit-serial reconfigurable VLSI based on multiple-valued current-mode circuits is proposed for highly parallel operations. In the multiple-valued logic, an arbitrary 2-variable binary function and full-adder Sum can be programmed in a common hardware resource, which leads to high utilization of the cell. Moreover, superposition of data and start signals in a single interconnection is introduced to reduce the number of interconnections and switches. In a 90nm CMOS design rule, the area and the area-time product of the proposed cell can be reduced to 78% and 92%, respectively, in comparison with those of an almost equivalent binary CMOS implementation.

S3L-4
Optimal Spot-checking to minimize the Computation Time in Volunteer Computing Systems
Kan Watanabe, Masaru Fukushi, Susumu Horiguchi

-- Abstract --

This paper proposes an optimization technique for spot-checking tominimize the computation time of volunteer computing (VC) systems with malicious participants who return erroneous results.Credibility based voting with spot-checking is one of the mostpromising approaches to guaranteeing the computational correctness of VC systems.In this method, there is an optimal value for the spot-check rate;that is, the increase of the spot-check rate reduces the necessaryredundancy for voting, which in turn increases the computation time due to the spot-checkingitself.From this trade-off, the estimation of the optimal spot-check rate istherefore the major concern to minimize the computation time of the credibility based voting. The key idea is to represent the mathematical expectation of thecomputation time as a function of the spot-check rate to obtain an estimate of the optimal spot-check rate before thecomputation.

S3L-5
VLSI Architecture for Hierarchical Temporal Memory
Wim J.C. Melis, Michitaka Kameyama

-- Abstract --

A large number of real world applications, such as imagerecognitionand understanding, can still not be performed easily byconventionalalgorithms in comparison with the human brain.Implementingapplications that require such intelligence, mighttherefore require adifferent approach, for which Hierarhical TemporalMemory (HTM) seemsa promising framework. Currently HTM exists as asoftware model.However, this software implementation has itsperformance limitationsand a distinct learning and operating mode.This paper proposes apossible architecture for VLSI implementationthat also allowslearning during normal operation.

Go back to the Top.

S4J. Nanofabrication and Molecular Electronics

Location: 3F Room J
Date & Time: 14:30-15:30, Friday, 17 October 2008
Chair: Masao Sakuraba, Associate Professor

S4J-1
SEM and EDX analyses on the growth of MoO₂ nanotube with rectangular cross sections
Hiroyuki Handa, Toshimi Abe, Maki Suemitsu

-- Abstract --

MoO₂ nanotube with rectangular cross sections is known to be formed with combustion-flame chemical vapor deposition. We have in this study investigated the growth behavior of the MoO₂ nanotubes in detail by using scanning electron microscopy (SEM) and energy-dispersive x-ray microanalysis (EDX). As a result, we found that the growth proceeds via a two-stage pathway: a bulk cluster growth and a following nanotube growth. A possible growth model will be presented.

S4J-2
Temperature oscillation during 3C-SiC growth on Si substrate by gas-source MBE
Eiji Saito, Maki Suemitsu

-- Abstract --

Temperature oscillation has been observed during gas-source molecular beam epitaxy (GSMBE) of 3C-SiC on Si substrate using monomethyl silane. The period of the “temperature oscillation” is shown to correspond to λ/2n, with the λ and n being the wavelength of the light used in the monochromatic optical pyrometry and the refractive index of the grown film, respectively. The decay of the oscillation is shown to be caused by roughening of the surface. Pyrometric interferomety may provide a feasible real-time monitoring method for the growth of 3C-SiC film on Si substrates.

S4J-3
Incubation-Free Growth of Polycrystalline Si on Polyethylene-Terephthalate (PET) Films using Pulsed-Plasma CVD under near Atmospheric Pressure
Mitsutaka Matsumoto, Yohei Inayoshi, Syougo Murashige, Maki Suemitsu, Setsuo Nakajima, Tsuyoshi Uehara, Yasutake Toyoshima

-- Abstract --

Polycrystalline-Si films grown on polyethylene-terephthalate(PET)substrates without incubation layers have been achieved with ahighgrowth rate (40 nm/min) using pulsed-plasma chemical vapordeposition(CVD) under near atmospheric pressure. Features ofthepolycrystalline-Si films were observed usingRaman-scatteringspectroscopy and cross-sectional transmission electronmicroscopy(X-TEM). Observations by X-TEM suggest that the Sipolycrystallitesare generated directly on the PET substrates withoutforming theincubation layers.

S4J-4
Measurement of lattice constants for diamond single crystal by repetition growth using the MPCVD process by X-ray diffractometer using the Bond method
Takashi Shishitani, Mototaka Arakawa, Yuji Ohashi, Jun-ichi Kushibiki, Yoshiaki Mokuno, Akiyoshi Chayahara, Shinichi Shikata

-- Abstract --

A microwave plasma CVD (MPCVD) is a possible method to grow a largediamond single crystal. A (100) multi-layered diamond specimen grownon a commercial (100) substrate grown by thehigh-temperature/high-pressure (HPHT) method was obtained with athickness of about 10 mm by 31-times repetition of high-rate epitaxialgrowth by the MPCVD method with nitrogen addition in methane andhydrogen source gases. A distribution of lattice constants wasmeasured by an X-ray diffractometer using the Bond method. The latticeconstants were influenced by nitrogen concentrations incorporatedduring the MPCVD process. Nitrogen concentrations were evaluated fromexperimental data for HPHT IIa and Ib single crystal substrates, wemeasured.

Go back to the Top.

S4K. Optical Network

Location: 3F Room K
Date & Time: 14:30-15:15, Friday, 17 October 2008
Chair: Masato Yoshida, Assistant Professor

S4K-1
A Nonblocking Self-Routing WDM Switch Based on Arrayed Waveguide Grating
Yusuke Fukushima, Xiaohong Jiang, Susumu Horiguchi

-- Abstract --

Arrayed waveguide grating (AWG) is a promising technology forconstructing high-speed large-capacity WDM switches, because itcan switch fast, is scalable to large size and consumes littlepower. To take the full advantages of high-speed AWG, the routingcontrol of a massive AWG-based switch should be as simple aspossible. In this paper, we focus on the self-routing design ofAWG-based switches with O(1) constant routing complexity andpropose a novel construction of self-routing AWG switches that canguarantee the attractive nonblocking property for both thewavelength-to-wavelength and wavelength-to-fiber request models.

S4K-2
Blocking and Delay Analysis of Single Wavelength Optical Buffer with General Packet Size Distribution
Jianming Liu, Tong Tony Lee, Xiaohong Jiang, Susumu Horiguchi

-- Abstract --

Buffers are essential components of any packet switch for resolving contentions among arriving packets. Optical buffers are composed of Fiber Delay Lines (FDL) and have finite time resolution property, which may lead to excess load and prolong packet delay. In this paper, a novel queueing model of optical buffer is proposed, the closed-form expressions of blocking probability and mean delay are derived to explore the trade-off between buffer performance and system parameters, such as the length of the optical buffer, the time granularity of FDLs, and to evaluate the overall impact of packet length distribution on the buffer performance.

S4K-3
Frequency division multiplexed 1-Gsymbol/s, 64 QAM coherent optical transmission with a spectral efficiency of 8.6 bit/s/Hz
Keisuke Kasai, Masato Yoshida, Toshihiko Hirooka, Masataka Nakazawa

-- Abstract --

Quadrature amplitude modulation (QAM) is an attractive modulation format for realizing optical communication systems with a high spectral efficiency of much greater than 1 bit/s/Hz. We have already proposed a QAM coherent optical transmission system by using heterodyne detection with a frequency-stabilized laser and an optical phase-locked loop and demonstrated　64 and 128 QAM coherent optical transmissions with a single channel. In this paper, we describe a three-channel frequency division multiplexed (FDM) and polarization-multiplexed 1-Gsymbol/s, 64 QAM coherent optical transmission over 160 km with a 1.4 GHz spacing. The spectral efficiency was reached as high as 8.6 bit/s/Hz.

Go back to the Top.

S4L. VLSI and FPGA

Location: 3F Room L
Date & Time: 14:30-15:30, Friday, 17 October 2008
Chair: Naofumi Homma, Assistant Professor

S4L-1
A Low-Power Field-Programmable VLSI Based on an Asynchronous Architecture
Shota Ishihara, Masanori Hariyama, Michitaka Kameyama

-- Abstract --

This paper presents a novel asynchronous architectureoffield-programmable gate arrays (FPGAs) to reduce thepowerconsumption. To reduce the dynamic power, we use thelevel-encodeddual-rail encoding which archives lowest-power indual-rail encodingand propose its area-efficient implementation. Toreduce the stand-bypower, we propose autonomous fine-grainpower-gating with smalloverhead at the control circuit. In a 90nm CMOSprocess, the dynamicpower and stand-by power of the proposedfield-programmable VLSI arerespectively reduced to 44% and 36%compared with a typicalsynchronous FPGA.

S4L-2
Nonvolatile Lookup Table Circuit for a Standby-Power-Free Field-Programmable Gate Array
Daisuke Suzuki, Takahiro Hanyu

-- Abstract --

A new lookup-table (LUT) circuit, where tunneling magneto-resistive (TMR) devices with MOS transistors are combined, is presented for a standby-power-free field-programmable gate array (FPGA).Since nonvolatile storage elements are distributed over a MOS-transistor logic network, the state transition between standby and active mode can be instantly changed with maintaining circuit-configuration information.The efficiency of the proposed LUT circuit is demonstrated using a simple arithmetic circuit in comparison with those a conventional LUT together with an external nonvolatile memory

S4L-3
Interacting Self-Timed Pipelines and Elementary Applications
Kazuhiro Komatsu, Shuji Sannomiya, Makoto Iwata, Suguru Kameda, Kazuo Tsubouchi

-- Abstract --

The self-timed (clockless) pipeline (STP) is one of the mostpromisingVLSI architectures, achieving efficient utilization ofabundanttransistors, ultra low power, and easy-to-design with signalintegrityand low EMI. This paper proposes a novel scheme ofinteracting STPs bywhich various pipelines distributed andinterconnected can realizehighly functional stream processing. Thepaper also proposes a set ofelementary coupling control modules thatrealize various combinationsof stream processing between pipelines,and then discusses thepracticality of the proposed scheme through theLSI design of someapplication modules.

S4L-4
Chosen-Message SPA Attacks against FPGA-Based RSA Hardware Implementations
Atsushi Miyamoto, Naofumi Homma, Takafumi Aoki, Akashi Satoh

-- Abstract --

This paper presents SPA (Simple Power Analysis) attacks against public-key cryptosystems implemented on an FPGA platform. The SPA attack investigates a power waveform generated by a cryptographic module, and reveals a secret key in the module. We focus on chosen-message SPA attacks, which enhances the differences of operating waveforms between multiplication and squaring correlated to the secret key by using the input of particular messages. In this paper, we implemented four-types of RSA processors on an FPGA platform in combination with two variants of the Montgomery multiplication algorithm and two different types of multipliers for SPA attacks experiments. Then we demonstrated effectiveness of various chosen-message attacks, and investigated the characteristics of the attacks depending on the hardware architectures.

Go back to the Top.