A novel signal processing technique using adaptive neural network algorithm is applied for the on-line detection of harmonic current components generated by nonlinear current loads in the single-phase diode bridge rectifier and it can efficiently determine the harmonic current components in real time. The validity of this active filtering processing system to compensate current harmonics is proved on the basis of simulation results.

A configurable device "PCA-Chip2" implements the concept of Plastic Cell Architecture, which is an extension of programmable logic devices. This paper presents basic design tools for the PCA-Chip2 as the first step to develop the total design environment. Given a C description of a target function, configuration data for PCA-Chip2 is automatically generated by the tools. Trial designs by the tools are also presented to demonstrate the practicability of the proposed approach.

MRAM-writing circuitry to compensate for the thermal variation of the magnetization-reversal current is proposed. The writing current of the proposed circuitry is designed to decrease in proportion to an increase in temperature. This technique prevents multiple-write failures from degrading 1 Gb MRAM yield where the standard deviation of magnetization-reversal current variation from other origins is less than 5%.

The label placement problem is one of the most important problems in geographic information systems, cartography, graph drawing and graphical interface design. In this paper, we consider the problem of labeling points and curves in maps drawn from digital data. In digital maps, a curve is represented as a set of points and consists of many small segments. The label for each curve must be placed alongside the corresponding curve. We define a continuous labeling space for points and curves, and present an algorithm using this space for positioning labels. Computational results for subway and JR train maps in Tokyo are presented.

This paper describes multigrain parallel processing on OSCAR (Optimally SCheduled Advanced multiprocessoR) chip multiprocessor architecture. OSCAR compiler cooperative chip multiprocessor architecture aims at development of scalable, high effective performance and cost effective chip multiprocessor with ease of use by compiler supports. OSCAR chip multiprocessor architecture integrates simple single issue processors having distributed shared data memory for optimal use of data locality over different loops and fine grain data transfer and synchronization, local data memory for private data recognized by compiler, and compiler controllable data transfer unit for overlapping data transfer to hide data transfer overhead. This OSCAR chip multiprocessor and OSCAR multigrain parallelizing compiler have been developed simultaneously. Performance of multigrain parallel processing on OSCAR chip multiprocessor architecture is evaluated using SPEC fp 2000/95 benchmark suite. When microSPARC like single issue core is used, OSCAR chip multiprocessor architecture gives us 2.36 times speedup in fpppp, 2.64 times in su2cor, 2.88 times in turb3d, 2.98 times in hydro2d, 3.84 times in tomcatv, 3.84 times in mgrid and 3.97 times in swim respectively for four processors against single processor.

We propose a new algorithm for blind source separation (BSS), in which frequency-domain independent component analysis (FDICA) and time-domain ICA (TDICA) are combined to achieve a superior source-separation performance under reverberant conditions. Generally speaking, conventional TDICA fails to separate source signals under heavily reverberant conditions because of the low convergence in the iterative learning of the inverse of the mixing system. On the other hand, the separation performance of conventional FDICA also degrades significantly because the independence assumption of narrow-band signals collapses when the number of subbands increases. In the proposed method, the separated signals of FDICA are regarded as the input signals for TDICA, and we can remove the residual crosstalk components of FDICA by using TDICA. The experimental results obtained under the reverberant condition reveal that the separation performance of the proposed method is superior to those of TDICA- and FDICA-based BSS methods.

This study is a detailed numerical investigation on the relations between the performance of the RZ format single-channel transmission, and the chromatic dispersion of transmission fiber and pre-compensation ratio. We observed the transition from the SPM dominant low dispersion region to the intra-channel nonlinearities dominant high dispersion region, and found that the EOP is very sensitive to the pre-compensation ratio when the dispersion assumes a intermediate value. Furthermore, by analyzing the optical power-dependence of the EOP and other nonlinear impairments, we found that the amplitude fluctuation resulting from IFWM is dominant in determining the EOP in the transmission systems employing highly dispersed pulses.

The Neural Network (NN) is applied to recognize basic PCB configurations using its magnetic near-field spectra and radiated far-field emission. The learning process is accomplished by using the computed spectra of the radiated field from PCBs having different configurations. The anomaly is detected through the monitoring of the spectra's amplitude frequency by injecting a voltage pulse at the PCB configuration. The trained NN is then applied to the identification of PCB layouts from radiated emission measurements. The trained NN can identify all of those PCB configurations from the magnetic near-field spectra and the radiated far-field EMI. Moreover, the calculated results of the NN are compared with the actual far-field measurements and other models for evaluation. Finally, the NN used for predicting far-field emission from their magnetic near-field measurement is proposed. Experiments show that the NN can predict the far-field spectra from the magnetic near-field spectra.

Compressed video bitstreams are very sensitive to transmission errors. If we lose packets or receive them with errors during transmission, not only the current frame will be corrupted, but also errors will propagate to succeeding frames. Therefore, we need various mechanisms to protect data and reduce the effects of transmission errors. Error concealment is a data recovery technique that enables the decoder to conceal the effects of transmission errors by predicting the lost or corrupted video data from the previously reconstructed error free information. Motion vector recovery and motion compensation with the estimated motion vector is a good approach to conceal the corrupted macroblock data. In this paper, we show that it is reasonable to use the estimated motion vector to conceal the lost macroblock by providing macroblock distortion models. After we propose a new motion vector recovery algorithm based on optical flow fields, we compare its performance to those of conventional error concealment methods.

Datacycle is an information delivery system designed in Bellcore . It uses broadcast delivery and multiple channels of an optical fiber to provide a large volume of information to many users. In this system, the mean access time depends on the broadcast delivery schedule because different information items (called pages) may have different popularity and there are multiple channels for concurrent broadcast delivery. In this paper, we design broadcast delivery schedules for M channels where M 2 and our objective is to minimize the mean access time. We show that this design problem can be divided into two subproblems: (1) divide the pages into M partitions such that the pages of each partition are broadcast in a distinct channel and (2) determine a broadcast schedule for the pages of each partition. We analyze and solve these subproblems, and we demonstrate that the schedules found can nearly reach a lower bound on the minimal mean access time.

In this paper, two low power hardware structures essential for MPEG-4 video codec are proposed for portable applications. First, an adaptive bit resolution control (ABRC) scheme is proposed for a processing element (PE) in a systolic-array type motion estimator (ME). By appropriately modifying the datapath of PE to exploit the correlations in pixel values, its structure is optimized in terms of both hardware cost and low power consumption. As a result, power is saved up to 29% compared with a conventional PE while the computation accuracy is preserved and the overhead is kept negligible. Second, a low power motion compensation (MC) accelerator is proposed. By embedding DRAM whose structure is optimized for low power consumption, the power consumption for external data I/Os is dramatically reduced. In addition, distributed nine-tiled block mapping (DNTBM) with partial activation scheme in the frame buffer reduces the power for accessing frame buffer up to 31% compared to a conventional 1-bank tiled mapping. With the proposed MC accelerator, MPEG-4 SP@L1 decoding system is fabricated using 0.18 µm embedded memory logic (EML) technology.

A wavelet-based vector quantization scheme for image compression is introduced here. The proposed scheme obtains a better compression efficiency by the following three methods. (1) Utilizing the correlation among wavelet coefficients. (2) Placing different emphasis on wavelet coefficients at different levels. (3) Preserving the most important information of the image. In our experiments, simulation results show that this technique outperforms the recent SMVQ-ABC [1] and WTC-NIVQ [2] techniques.

Caching web files reduces user response time as well as network traffic. When implementing caches, the file caching problem must be addressed; the problem is how to determine which files should be evicted from a cache when there is insufficient space for storing a new file so that the sum of the mis-hit (fault) file costs is minimized. Greedy-Dual-Size (GDS) is the best online algorithm in terms of competitiveness, i. e. , (k)/(k-h+1)-competitive, where k and h are the storage space of, respectively, GDS and an optimal offline algorithm. GDS performs very well even in trace-driven simulations. The worst-case time taken to service a request is another important measure for online file caching algorithms since slow response times render caching meaningless from the client's view point. This paper proposes a fast randomized (k)/(k-h+1)-competitive algorithm that performs in O(2log ^* k) time per file eviction or insertion, whereas GDS takes O(log k) time, where 2log ^* k is a much slower increasing function than log k. To confirm its practicality, we conduct trace driven simulations. Experimental results show that our algorithm attains only slightly worse byte hit rates and sufficiently large reduced latency in comparison with GDS, and our algorithm is a good candidate for caches requiring high-speed processing such as second-level caches in the large networks.

A heterogeneous parallel computing environment consisting of different types of workstations and communication links plays an important role in parallel computing. In many applications on the system, collective communication operations are commonly used as communication primitives. Thus, design of the efficient collective communication operations is the key to achieve high-performance parallel computing. But the heterogeneity of the system complicates the design. In this paper, we consider design of an efficient gather operation, one of the most important collective operations. We show that an optimal gather schedule is found in O(n2k-1) time for the heterogeneous parallel computing environment with n processors of k distinct types, and that a nearly-optimal schedule is found in O(n) time if k=2.

This paper presents a dynamic node decaying method (DNDM) for layered artificial neural networks that is suitable for classification problems. Our purpose is not to minimize the total output error but to obtain high generalization ability with minimal structure. Users of the conventional back propagation (BP) learning algorithm can convert their program to the DNDM by simply inserting a few lines. This method is an extension of a previously proposed method to more general classification problems, and its validity is tested with recent standard benchmark problems. In addition, we analyzed the training process and the effects of various parameters. In the method, nodes in a layer compete for survival in an automatic process that uses a criterion. Relatively less important nodes are decayed gradually during BP learning while more important ones play larger roles until the best performance under given conditions is achieved. The criterion evaluates each node by its total influence on progress toward the upper layer, and it is used as the index for dynamic competitive decaying. Two additional criteria are used: Generalization Loss to measure over-fitting and Learning Progress to stop training. Determination of these criteria requires a few human interventions. We have applied this algorithm to several standard benchmark problems such as cancer, diabetes, heart disease, glass, and iris problems. The results show the effectiveness of the method. The classification error and size of the generated networks are comparable to those obtained by other methods that generally require larger modification, or complete rewriting, of the program from the conventional BP algorithm.

This paper concerns recognizing 3-dimensional object using proposed multi-layer block model. In particular, we aim to achieve desirable recognition performance while restricting the computational load to a low level using 3-step feature extraction procedure. An input image is first precisely partitioned into hierarchical layers of blocks in the form of base blocks and overlapping blocks. The hierarchical blocks are merged into a matrix, with which abundant local feature information can be obtained. The local features extracted are then employed by the kernel based support vector machines in tournament for enhanced system recognition performance while keeping it to low dimensional feature space. The simulation results show that the proposed feature extraction method reduces the computational load by over 80% and preserves the stable recognition rate from varying illumination and noise conditions.

This paper proposes a silence compression algorithm operating at multi-rates (MR) and with dual-bandwidths (DB), a narrowband and a wideband, for the MPEG (Moving Picture Experts Group)-4 CELP (Code Excited Linear Prediction) standard. The MR/DB operations are implemented by a Variable-Frame-size/Dual-Bandwidth Voice Activity Detection (VF/DB-VAD) module with bandwidth conversions of the input signal, and a Variable-Frame-size Comfort Noise Generator (VF-CNG) module. The CNG module adaptively smoothes the Root Mean Square (RMS) value of the input signal to improve the coding quality during transition periods. The algorithm also employs a Dual-Rate Discontinuous Transmission (DR-DTX) module to reduce an average transmission bitrate during silence periods. Subjective test results show that the proposed silence compression algorithm gives no degradation in coding quality for clean and noisy speech signals. These signals include about 20 to 30% non-speech frames and the average transmission bitrates are reduced by 20 to 40%. The proposed algorithm has been adopted as a part of the ISO/IEC MPEG-4 CELP version 2 standard.

In this paper, we investigate the electron-hole energy states and energy gap in three-dimensional (3D) InAs/GaAs quantum rings and dots with different shapes under external magnetic fields. Our realistic model formulation includes: (i) the effective mass Hamiltonian in non-parabolic approximation for electrons, (ii) the effective mass Hamiltonian in parabolic approximation for holes, (iii) the position- and energy-dependent quasi-particle effective mass approximation for electrons, (iv) the finite hard wall confinement potential, and (v) the Ben Daniel-Duke boundary conditions. To solve the 3D nonlinear problem without any fitting parameters, we have applied the nonlinear iterative method to obtain self-consistent solutions. Due to the penetration of applied magnetic fields into torus ring region, for ellipsoidal- and rectangular-shaped quantum rings we find nonperiodical oscillations of the energy gap between the lowest electron and hole states as a function of external magnetic fields. The nonperiodical oscillation is different from 1D periodical argument and strongly dependent on structure shape and size. The result is useful to study magneto-optical properties of the nanoscale quantum rings and dots.

Blind separation of sources from their linear mixtures is a well understood problem. However, if the mixtures are nonlinear, this problem becomes generally very difficult. This is because both the nonlinear mapping and the underlying sources must be learned from the data in a blind manner, and the problem is highly ill-posed without a suitable regularization. In our approach, multilayer perceptrons are used as nonlinear generative models for the data, and variational Bayesian (ensemble) learning is applied for finding the sources. The variational Bayesian technique automatically provides a reasonable regularization of the nonlinear blind separation problem. In this paper, we first consider a static nonlinear mixing model, with a successful application to real-world speech data compression. Then we discuss extraction of sources from nonlinear dynamic processes, and detection of abrupt changes in the process dynamics. In a difficult test problem with chaotic data, our approach clearly outperforms currently available nonlinear prediction and change detection techniques. The proposed methods are computationally demanding, but they can be applied to blind nonlinear problems of higher dimensions than other existing approaches.

Changing vehicle structures and backgrounds makes it very difficult to correctly extract a license plate region from a vehicle image. In this paper, we propose a simple method to extract the license plate region using edge properties of wavelet subband. The High Frequency Subband (HFS) of an image has edge information for each direction. Edge information is concentrated in each direction of the Headlight-Radiator-Headlight (H-R-H) and the license plate region compared to other regions in the vehicle image. This paper shows a license plate region extraction method using these edge properties and our experimental results with various vehicle images.