Shinya FUKUMOTO Noritaka SHIGEI Michiharu MAEDA Hiromi MIYAJIMA
Neural networks for Vector Quantization (VQ) such as K-means, Neural-Gas (NG) network and Kohonen's Self-Organizing Map (SOM) have been proposed. K-means, which is a "hard-max" approach, converges very fast. The method, however, devotes itself to local search, and it easily falls into local minima. On the other hand, the NG and SOM methods, which are "soft-max" approaches, are good at the global search ability. Though NG and SOM exhibit better performance in coming close to the optimum than that of K-means, the methods converge slower than K-means. In order to the disadvantages that exist when K-means, NG and SOM are used individually, this paper proposes hybrid methods such as NG-K, SOM-K and SOM-NG. NG-K performs NG adaptation during short period of time early in the learning process, and then the method performs K-means adaptation in the rest of the process. SOM-K and SOM-NG are similar as NG-K. From numerical simulations including an image compression problem, NG-K and SOM-K exhibit better performance than other methods.
Self-organizing map is a widely used tool in high-dimensional data visualization. However, despite its benefits of plotting very high-dimensional data on a low-dimensional grid, browsing and understanding the meaning of a trained map turn to be a difficult task -- specially when number of nodes or the size of data increases. Though there are some well-known techniques to visualize SOMs, they mainly deals with cluster boundaries and they fail to consider raw information available in original data in browsing SOMs. In this paper, we propose our Factor controlled Hierarchical SOM that enables us select number of data to train and label a particular map based on a pre-defined factor and provides consistent hierarchical SOM browsing.
This paper describes a method of analyzing musical sound using a self-organizing map. To take compound factors into account, energy spectra whose frequency ranges were based on the psycho-acoustic experiments were used as input data. Results for music compact discs confirmed that our method could effectively display the positioning and relationships among musical sounds on a map.
Mitsuhiro SHIMOZAWA Takatoshi KATSURA Kenichi MAEDA Eiji TANIGUCHI Takayuki IKUSHIMA Noriharu SUEMATSU Kenji ITOH Yoji ISOTA Tadashi TAKAGI
This paper presents an even harmonic mixer using self-biased anti-parallel diode pair (APDP). A proposed self-biased APDP is composed of a pair of diodes and self-bias series resistors. At high LO injection level, rectified current is generated by the diodes and reverse voltage is applied to the diodes by the self-bias resistor. Therefore, rapid degradation of conversion loss at high LO input level can be avoided. The effect of self-bias resistor is explained by using simplified behavior model and harmonic balance method, and is evaluated by the measurements of an L-band even harmonic type direct conversion mixer.
Masaaki KUBO Zaher AGHBARI Kun Seok OH Akifumi MAKINOUCHI
This paper proposes a technique for indexing, clustering and retrieving images based on their edge features. In this technique, images are decomposed into several frequency bands using the Haar wavelet transform. From the one-level decomposition sub-bands an edge image is formed. Next, the higher order auto-correlation function is applied on the edge image to extract the edge features. These higher order autocorrelation features are normalized to generate a compact feature vector, which is invariant to shift, image size. We used direction cosine as measure of distance not to be influenced by difference of each image's luminance. Then, these feature vectors are clustered by a self-organizing map (SOM) based on their edge feature similarity. The performed experiments show higher precision and recall of this technique than traditional ways in clustering and retrieving images in a large image database environment.
Sebastien NUTTINCK Edward GEBARA Stephane PINEL Joy LASKAR
We report the investigation of major dispersion mechanisms such as self-heating, trapping, current collapse, and floating-body effects present in AlGaN/GaN HFETs. These effects are analyzed using DC/Pulsed IV, load-pull, low-frequency noise systems, and a cryogenic probe station. This study leads to a better understanding of the device physics, which is critical for accurate large-signal modeling and device optimization.
Shinji NIWA Hiroshi YASUKAWA Ichi TAKUMI Masayasu HATA
The tectonic activities that precede significant earthquakes release electromagnetic (EM) waves that can be used as earthquake precursors. We have been observing EM radiation in the ELF (extremely low frequency) band at about 40 observation stations in Japan for predicting significant earthquakes. The recorded signals contain, however, several noise components generated from the ionosphere, human activity, and so on. Most background noise in observed signal is attributed to lightning in the tropics. This paper proposes method based on PCA (principal component analysis) to extract signals from large data sets. The good performance of the proposed method is confirmed.
This paper focuses on a global ultrasonic system for self-localization of a mobile robot. The global ultrasonic system consists of some ultrasonic generators fixed at some arbitrary position in the global coordinates and two receivers in the moving coordinates of the mobile robot. This system is used to obtain the state vector of the mobile robot in the global coordinates from the distance measurement between the ultrasonic generator and the receiver. In order to avoid the cross-talk and to synchronize the ultrasonic sensors, the sequential cuing technique using small-sized radio frequency module is adopted. An extended Kalman filter algorithm is used to process the noisy ultrasonic signal and to estimate the state vector. Computer simulations and experiments are conducted to verify the effectiveness of the proposed global ultrasonic system.
This paper describes an analysis of the electromagnetic interference (EMI) aspects of electrostatic discharge (ESD), which sometimes causes serious damage to electrical systems. To classify EMI-related properties resulting from ESD events, we used a self-organizing neural network, which can map high-dimensional data into simple geometric relationships on a low-dimensional display. Also, to clarify the effect of a high-speed moving discharge, we generated one-shot discharges repeatedly and measured the ESD current in the time domain to obtain its EMI-related characteristics of this phenomenon. Based on the measured data, we examined several differential properties of ESD waveforms including the maximum amplitude and energy level, and analyzed these multi-dimensional data using the self-organizing neural network scheme. The results showed that the high-speed moving discharges can increase the maximum amplitude, relative energy, and entropy of ESD events, and that the positioning of the EMI level of each ESD event can be effectively visualized in a two-dimensional space.
Yozo SHOJI Kiyoshi HAMAGUCHI Hiroyo OGAWA
We describe a low-cost and extremely stable millimeter-wave transmission system that uses a double-side-band (DSB) millimeter-wave self-heterodyne transmission technique. This technique allows us to use a comparatively low-cost and unstable millimeter-wave oscillator regardless of the modulation format. Furthermore, a transmission band-pass-filter (BPF) is not needed in the millimeter-wave band. The system cost can therefore be substantially reduced. We have theoretically and experimentally evaluated the carrier-to-noise power ratio (CNR) performance that can be obtained when using this technique relative to that attainable through a conventional millimeter-wave self-heterodyne technique where a single-side-band signal is transmitted. Our results show that the DSB self-heterodyne transmission technique can improve CNR by more than 3 dB.
Ki Chan EUN Young Chul LEE Byung Gun CHOI Dae Jun KIM Chul Soon PARK
Fully embedded spiral inductors in a low loss dielectric multi-layer were designed and fabricated using a low temperature co-fired ceramics (LTCC) technology for RF SIP (system in package) integrations. The line width/space and the number of spiral layers were optimized within five layers of LTCC dielectric for high Q-factor, high self-resonant frequency (SRF), process easiness, and compact size. The embedded multi-layer spiral inductors reveal better performance in terms of Q-factor, SRF and the effective inductance Leff than planar spiral inductors of the same dimension and number of turns. The optimized multi-layer spiral inductor shows maximum Q of 56, Leff of 6.6 nH at Qmax and SRF of 3.6 GHz while planar spiral inductors have maximum Q of 49, Leff of 5.8 nH at Qmax and SRF of 3.0 GHz.
Soo-Hyun OH Masahiro MAMBO Hiroki SHIZUYA Dong-Ho WON
In 1991 Girault proposed a key agreement protocol based on his new idea of self-certified public key. Later Rueppel and Oorschot showed variants of the Girault scheme. All of these key agreement protocols inherit positive features of self-certified public key so that they can provide higher security and smaller communication overhead than key agreement protocols not based on self-certified public key. Even with such novel features, rigorous security of these protocols has not been made clear yet. In this paper, we give rigorous security analysis of the original and variants of Girault key agreement protocol under several kinds of active attacker models. In particular we show that protocols are either insecure or proven as secure as the Diffie-Hellman problem over Zn with respect to the reduction among functions of computing them. Analyzed protocols include a new variant of 1-pass protocol. As opposed to the original 1-pass protocol, the new variant provides mutual implicit key authentication without increasing the number of passes.
Tsunehiro YOSHINAGA Katsushi INOUE
This paper investigates the accepting powers of deterministic, Las Vegas, self-verifying nondeterministic, and nondeterministic one-way multi-counter automata with time-bounds. We show that (1) for each k1, there is a language accepted by a Las Vegas one-way k-counter automaton operating in real time, but not accepted by any deterministic one-way k-counter automaton operating in linear time, (2) there is a language accepted by a self-verifying nondeterministic one-way 2-counter automaton operating in real time, but not accepted by any Las Vegas one-way multi-counter automaton operating in polynomial time, (3) there is a language accepted by a self-verifying nondeterministic one-way 1-counter automaton operating in real time, but not accepted by any deterministic one-way multi-counter automaton operating in polynomial time, and (4) there is a language accepted by a nondeterministic one-way 1-counter automaton operating in real time, but not accepted by any self-verifying nondeterministic one-way multi-counter automaton operating in polynomial time.
Michael LOGOTHETIS Ioannis NIKOLAOU
Modern network technologies gave rise to intelligent network reconfiguration schemes for restoration purposes and several network self-healing schemes, exploiting the capabilities of network elements (NE), have already been proposed. Each self-healing scheme has its own characteristics, regarding restoration time, flexibility, restoration cost and exploitation of NEs. Integrated self-healing networks, which combine more than one survivability techniques, mainly the Shared Self-Healing Rings (SSR) with the Dynamic Self-Healing Networks (DSN), can achieve higher network survivability and cost-effective network design. In this paper, we propose two algorithms for the design of spare and working channel capacities for integrated self-healing networks. In the first algorithm, A1, we do not take into account the capacity of network nodes, while in the second algorithm, A2, we take into account the limited capacity of network nodes. These algorithms are based on the shortest path principles, similarly to a previous algorithm (old algorithm) proposed by scientists of NEC Corporation for integrated self-healing network design. By the new algorithms we achieve more savings than by the old algorithm in total network capacity. On the other hand, strong motivation for the development of the new algorithms is the fact that the procedural steps of the old algorithm are not homogeneous; the old algorithm incorporates both heuristics and analytical methods, in contrast to the new algorithms that are pure heuristics. Moreover, we introduce restrictions in node-capacities of the network that they were not included in the old algorithm.
Invention and development of the Yagi-Uda antenna, and the self-complementary antenna are described. Analysis methods of large loop antennas and the improved circuit theory (ICT) for design of linear antennas are presented. Recent developments of axial mode helical antennas and spiral antennas for radiating circularly polarized waves are also described.
Fabian M. BUFLER Christoph ZECHNER Andreas SCHENK Wolfgang FICHTNER
The validity and capability of an iterative coupling scheme between single-particle frozen-field Monte Carlo simulations and nonlinear Poisson solutions for achieving self-consistency is investigated. For this purpose, a realistic 0.1 µm lightly-doped-drain (LDD) n-MOSFET with a maximum doping level of about 2.5 1020 cm-3 is simulated. It is found that taking the drift-diffusion (DD) or the hydrodynamic (HD) model as initial simulation leads to the same Monte Carlo result for the drain current. This shows that different electron densities taken either from a DD or a HD simulation in the bulk region, which is never visited by Monte Carlo electrons, have a negligible influence on the solution of the Poisson equation. For the device investigated about ten iterations are necessary to reach the stationary state after which gathering of cumulative averages can begin. Together with the absence of stability problems at high doping levels this makes the self-consistent single-particle approach (SPARTA) a robust and efficient method for the simulation of nanoscale MOSFETs where quasi-ballistic transport is crucial for the on-current.
Toshiyuki UMEDA Shoji OTAKA Kenji KOJIMA Tetsuro ITAKURA
This paper describes a low-power-supply 2-GHz CMOS up-converter. A current-mode mixing method using current adding and self-switching mixers is proposed for 1-V operation. The current-mode up-converter achieves conversion gain of 6.7 dB and linearity of 6.5-dBm OIP3 at 1 V. Balanced configuration and DC offset canceller reduce LO leakage below -40 dBc even with 20-mV Vth mismatches. The bias circuit of the IC is designed to maintain constant conversion gain for variation of temperature for practical usage. The measurement results indicate the proposed up-converter is applicable for future wireless systems.
Xiaotong HU Makoto FUJIMURA Yoko MAEMURA Hideo KURODA
In fractal image coding, for each range block, the best matching domain block is identified, and information from the best matching domains and range blocks are transmitted to the decoder for image reconstruction. In this paper, the similarity between range blocks and domain blocks is evaluated according to their centers of gravity. The number of searched domain blocks are reduced by limiting the candidates for the best matching domain blocks to those domain blocks whose similarity to the range block are high. Using simulation experiments, the number of candidates for the best matching domain blocks were reduced to about 10-23% of the current method. Thus, our proposed method had significantly reduced the number of searched domain blocks below the current method and at the same time it turns out that degradation of the reconstructed image was seldom observed.
Masashi NAKATSUGAWA Masahiro MURAGUCHI Yo YAMAGUCHI
We propose linearization techniques for MMIC amplifiers. The key points of these techniques are increased linearity of a newly-developed low-distortion MESFET (LD-FET) and maximized IP3 by combining the LD-FET with a high-gain depletion-mode MESFET (D-FET) with no increase in power consumption. The LD-FET is characterized by its unique channel dopant-profile prepared by a buried p-type ion-implantation and double n-type ion-implantations with high- and low-acceleration energies. This FET achieves flatter behavior in terms of mutual conductance (gm) compared with conventional MESFETs irrespective of changes in the gate bias voltage (Vgs). A self-alignment/selective ion-implantation process enables the LD-FET and D-FET to be fabricated simultaneously. This process encourages IP3 maximization of the multi-stage amplifier by appropriately combining the advantages of the two differently characterized MESFETs. We fabricated and tested a highly linearized two-stage MMIC amplifier utilizing the proposed techniques, and found that its third-order intermodulation ratio (IMR) performance was 8.7 dB better than that of conventional MMIC amplifiers at an input signal level of -20 dBm with no increase in current dissipation. The configuration constructed by using the proposed techniques equivalently reduces the current dissipation of the second stage to 1/2.72 times that of the conventional configuration, which requires a 2.72 times larger D-FET at the second stage to obtain an 8.7-dB IMR improvement. Furthermore, we were able to improve the IMR by 3.5 dB by optimizing the gate bias conditions for the LD-FET. These results confirm the validity of the proposed techniques.
Newaz M. S. RAHIM Takashi YAHAGI
Finite-state vector quantization (FSVQ) is a well-known block encoding technique for digital image compression at low bit rate application. In this paper, an improved feature map finite-state vector quantization (IFMFSVQ) algorithm using three-sided side-match prediction is proposed for image coding. The new three-sided side-match improves the prediction quality of input blocks. Precoded blocks are used to alleviate the error propagation of side-match. An edge threshold is used to classify the blocks into nonedge or edge blocks to improve bit rate performance. Furthermore, an adaptive method is also obtained. Experimental results reveal that the new IFMFSVQ reduces bit rate significantly maintaining the same subjective quality, as compared to the basic FMFSVQ method.