Parallel processing is a well-known approach to enhance the performance of communication subsystems. The several forms of parallelism embedded in communication protocols have been applied to the OSI protocol stack. However, the OSI protocol stack involves sequential processing due to the layered architecture. Thus, all the layers have been prevented from performing immediate processing as soon as the data arrives. To solve the problem, we apply a Multiple Instruction Single Data (MISD) parallel scheme to OSI processing for the network layer through the presentation layer. In the MISD scheme, different processors can be allocated to different layers and concurrently run the code for each layer. In contrast, the conventional approach adopts for a pipeline scheme that all the layers can be assigned to different pipeline stages and will be performed in a time interval and their dependence. The implementations have been performed to compare the pipeline scheme with the MISD scheme on the Parsytec Super Cluster consisting of 64 Transputers. The measures show that the MISD scheme has performance improvement as high as about 84% in comparison with the pipeline one.

The properties of the optical-microwave mixing process are investigated in detail. To describe these processes a new approach, the parametric method is introduced which provides a better description of the mixing phenomenon. The paper presents new experimental results on and new theoretical analysis methods for the optical-microwave mixing process and also for its dynamic behavior. The dynamic properties are very important in many applications when the light is intensity modulated by a high frequency or high bit rate signal. A remarkable decrease is observed in the mixing product with increasing optical modulation frequency. There are two reasons for it: the time constant exhibited by the depletion region between the substrate and the epitaxial layer and the optically induced substrate current which is increasing with the modulation frequency and doesn't contribute to the mixing effect. Understanding the optical-microwave mixing process provides new solutions for many applications. For example the optical-microwave mixing techniques offers several advantages in case of optical reception. In the detection process the modulation signal can be transposed to an intermediate frequency band (instead of the baseband) making possible a lower noise reception in a wider band. Another important and advantageous application is in the reception of subcarrier modulated optical signals.

A voltage-mode resonant forward converter with capacitor-input filter is proposed, and its static and dynamic characteristics for both half-wave type and full-wave type are revealed by analysis and experiment. As a result, this converter has prominent features of simplicity of circuit configuration, isolation between input and output and high stability.

Semi-Orthogonally Associative Memory neural network model (SAM) uses the orthogonal vectors in Un = {-1, 1}n as its characteristic patterns. It is necessary to select the optimum characteristic parameter n so as to increase the efficiency of this model used. This paper investigates the dynamic behavior and error correcting capability of SAM by statistical neurodynamics, and demonstrates that there exists a convergence criterion in tis recalling processes. And then, making use of these results, its optimum characteristic parameter is deduced. It is proved that, in the statistical sense, its recalling outputs converge to the desired pattern when the initial similar probability is larger than the convergence criterion and not true otherwise. For a SAM with N neurons, when its characteristic parameter is optimum, its memory capacity is N/2 ln ln N, the information storage capacity per connection weight is larger than 9/23 (bits/weight) and the radius of attractive basin of non-spurious stable state is about 0.25N. Computer simulations are done on this model and the simulation results are consistent with the results of theoretical analyses.

Three primary wiring capacitance components for multi-layer interconnect structure in sub-micron LSI were analyzed by using 2D/3D simulators, and an influence of neighboring wiring was investigated as a three-body problem. The investigated neighboring wiring are three kinds, and they are same-layer, upper-layer and under-layer wiring. An analytical model of each capacitance component was proposed for LPE (Layout Parameter Extraction) system, and its accuracy and application limit were discussed. This new model can estimate each capacitance component of complicated interconnect structure within 20% error.

CAMs (Content Addressable Memories) are functional memories which have functions such as word-parallel equivalence search, bilateral 1-bit data shifting between consecutive words, and word-parallel writing. Since CAMs can be integrated because of their regular structure, massively parallel CAM functions can be executed. Taking advantage of CAMs, Ishiura and Yajima have proposed a parallel fault simulation algorithm using a CAM. This algorithm, however, requires a large amount of CAM storage to simulate large-scale circuits. In this paper, we propose a new massively parallel fault simulation algorithm requiring less CAM storage, and compare it with Ishiura and Yajima's algorithm. Experimental results of the algorithm on CHARGE --the CAM-based hardware engine developed in our laboratory--are also reported.

We have developed a design technique for static logic circuits. Using this technique, we designed 1/2 divider-type 1:4 demultiplexer (DEMUX) and 2:1 selector-type 4:1 multiplexer (MUX) circuits, each of which is a key component in high-speed data multiplexing and demultiplexing. These circuits consist of double rail flip-flops (DR F/F). These flip-flops have a smaller mean internal capacitance than single rail flip-flops, making them suitable for high-speed operation. The DR F/F has a symmetric structure, so the double rail toggle flip-flop can put out an exactly balanced CK/CKN signal, which boosts the speed of the data flip-flops. The double rail structure enables 30% faster operation but consumes only 17% more power (per GHz) than a single rail circuit. In addition, our 0.25-µm process technology provides a 70% higher frequency operation than 0.5-µm process technology. At the supply voltage of 2.2 V, the DEMUX circuit and the MUX circuit operate at 4.55 GHz and 2.98 GHz, respectively. In addition, the 0.25-µm DEMUX circuit and the MUX circuit respectively consume 6.0 mW/GHz and 13.7 mW/GHz (＠1.3 V), which are only 12% of the power consumed by 3.3-V 0.5-µm circuits. Because of its high-speed and low-power characteristics, our design technique will greatly contribute to the progress of large-scale high-speed telecommunication systems.

This paper presents a new method for solving the structure-from-motion problem for optical flow. The fact that the structure-from-motion problem can be simplified by using the linearization technique is well known. However, it has been pointed out that the linearization technique reduces the accuracy of the computation. In this paper, we overcome this disadvantage by correcting the linearized solution in a statistically optimal way. Computer simulation experiments show that our method yields an unbiased estimator of the motion parameters which almost attains the theoretical bound on accuracy. Our method also enables us to evaluate the reliability of the reconstructed structure in the form of the covariance matrix. Real-image experiments are conducted to demonstrate the effectiveness of our method.

In this paper, we present a vision system for a depalletizing robot which recognizes carton objects. The algorithm consists of the extraction of object candidates and a labeling process to determine whether or not they actually exist. We consider this labeling a combinatorial optimization of labels, we propose a new labeling method applying Genetic Algorithm (GA). GA is an effective optimization method, but it has been inapplicable to real industrial systems because of its processing time and difficulty of finding the global optimum solution. We have solved these problems by using the following guidelines for designing GA: (1) encoding high-level information to chromosomes, such as the existence of object candidates; (2) proposing effective coding method and genetic operations based on the building block hypothesis; and (3) preparing a support procedure in the vision system for compensating for the mis-recognition caused by the pseudo optimum solution in labeling. Here, the hypothesis says that a better solution can be generated by combining parts of good solutions. In our problem, it is expected that a global desirable image interpretation can be obtained by combining subimages interpreted consistently. Through real image experiments, we have proven that the reliability of the vision system we have proposed is more than 98% and the recognition speed is 5 seconds/image, which is practical enough for the real-time robot task.

In order to evaluate quantitatively TMJ sound, TMJ sound in normal subject group, CMD patient group A with palpable sounds unknown to them, CMD patient group B with palpable sounds known to them, and CMD patient group C with audible sounds were detected by a contact microphone, and frequency analysis of the power spectra was performed. The power spectra of TMJ sound of normal subject group and patient group A showed patterns with frequency values below 100 Hz, whereas the power spectra of patient groups B and C showed distinctively different patterns with peaks of frequency component exceeding 100 Hz. As regards the cumulative frequency value, the patterns for each group clearly differed from those of other groups; in particular the 80% cumulative frequency value showed the greatest difference. From these results, it is assumed that the 80% cumulative frequency value can be used as an effective indicator for quantitative evaluation of TMJ sound.

Data reduction method for the earth-satellite-earth laser long-path absorption measurements of atmospheric trace species using the Retroreflector in Space (RIS) on the Advanced Earth Observing Satellite (ADEOS) is described. In the RIS experiment, atmospheric absorption will be measured with single-longitudinal-mode pulsed CO2 lasers and their second and third harmonics. High-resolution absorption spectra are measured by using the Doppler shift of the return beam which is caused by the satellite movement. Vertical profiles of O3 and CH4 are retrieved from the measured absorption line shapes with the inversion method. Also, column contents of CFC12, HNO3, CO2, CO, N2O are derived by the least squares method with assumptions on the relative vertical profiles. Errors in the measurement were evaluated by computer simulation.

In circuit partitioning for FPGAs, partitioned signal nets are connected using I/O blocks, through which signals are coming from or going to external pins. However, the number of I/O blocks per chip is relatively small compared with the number of logic-blocks, which realize logic functions, accommodated in the FPGA chip. Because of the I/O block limitation, the size of a circuit implemented on each FPGA chip is usually small, which leads to a serious decrease of logic-block utilization. It is required to utilize unused logic-blocks in terms of reducing the number of I/O blocks and realize circuits on given FPGA chips. In this paper, we propose an algorithm which partitions an initial circuit into multi-FPGA chips. The algorithm is based on recursive bi-partitioning of a circuit. In each bi-partitioning, it searches a partitioning position of a circuit such that each of partitioned subcircuits is accommodated in each FPGA chip with making the number of signal nets between chips as small as possible. Such bi-partitioning is achieved by computing a minimum cut repeatedly applying a network flow technique, and replicating logic-blocks appropriately. Since a set of logic-blocks assigned to each chip is computed separately, logic-blocks to be replicated are naturally determined. This means that the algorithm makes good use of unused logic-blocks from the viewpoint of reducing the number of signal nets between chips, i.e. the number of required I/O blocks. The algorithm has been implemented and applied to MCNC PARTITIONING 93 benchmark circuits. The experimental results demonstrate that it decreases the maximum number of I/O blocks per chip by a maximum of 49% compared with conventional algorithms.

There have been many developments on neural network research, and ability of a multi-layered network for classification of multi-spectral image data has been studied. We can classify non-Gaussian distributed data using the neural network trained by a back-propagation method (BPM) because it is independent of noise conditions. The BPM is a supervised classifier, so that we can get a high classification accuracy by using the method, so long as we can choose the good training data set. However, the multi-spectral data have many kinds of category information in a pixel because of its pixel resolution of the sensor. The data should be separated in many clusters even if they belong to a same class. Therefore, it is difficult to choose the good training data set which extract the characteristics of the class. Up to now, the researchers have chosen the training data set by random sampling from the input data. To overcome the problem, a hybrid pattern classification system using BPM and Kohonens feature mapping (KFM) has been proposed recently. The system performed choosing the training data set from the result of rough classification using KFM. However, how the remotely sensed data had been influenced by the KFM has not been demonstrated quantitatively. In this paper, we propose a new approach using the competitive weight vectors as the training data set, because we consider that a competitive unit represents a small cluster of the input patterns. The approach makes the training data set choice work easier than the usual one, because the KFM can automatically self-organize a topological relation among the target image patterns on a competitive plane. We demonstrate that the representative of the competitive units by principal component analysis (PCA). We also illustrate that the approach improves the classification accuracy by applying it on the classification of the real remotely sensed data.

A continuous-wave ultrasonic Doppler system using wide field ultrasound transducers was applied to telemeter blood velocity from the carotid artery of exercising subjects. Velocity spectrogram was obtained by Hanning windowed fast Fourier transformation of the telemetered data. Distortion caused by a high-pass filter and transducers in the telemetry system was discussed in the paper. As the maximum Reynolds number in our experiment was 1478 which is smaller than the critical level of 2000, the blood flow should be laminar. Spatial velocity profiles were then reconstructed from the velocity spectrogram. In this paper, we defined a converging index Q of the velocity spectrum to measure the bluntness of the spatial velocity distribution across the blood vessel. Greater Q, the blunter the velocity profile will be. Simulation results for spatial velocity distributions of theoretical parabolic flow and Gaussian-distribution spectra with varied Q value showed that the cut-off effect by a high-pass filter of cut-off frequency fc=200Hz in our system could be ignored when the axial velocity is larger than 0.30 m/s and Q is greater than 2.0. Our experimental results, in contrast to those obtained from phantom systems by us and by Hein and O'Brien, indicate that the distribution of blood velocity is much blunter than previously thought. The Q index exceeded 10 during systole, whereas it was 0.5 in parabolic flow. The peak of Q index lagged behind that of axial blood velocity by approximately 0.02s. The phase delay of the Q index curve might be due to the time needed for the red blood cells to form the non-homogeneous distribution.

This paper presents a new approach of data pipelining for mincut partitioning acceleration using a parallel computer. When using a parallel computer, it is important to have many processors always active, also the quality of the partitioning must not be sacrificed. Out approach covers both speed and quality. We choose the hardware CAD accelerator TP5000 to implement our approach, which consists of dedicated Very Long Instruction Word (VLIW) processors with high-speed interconnections. The TP5000 allows its connections to be reconfigured to optimize the data pipelines. We estimate that the speed of our approach using 10 processors on the TP5000 is 30 times faster than a SPARCStation-10.

The family Pk of graphs with proper-path-width at most k is minor-closed. It is known that the number of minimal forbidden minors for a minor-closed family of graphs is finite, but we have few such families for which all the minimal forbidden minors are listed. Although the minimal acyclic forbidden minors are characterized for Pk, all the minimal forbidden minors are known only for P1. This paper lists 36 minimal forbidden minors for P2, and shows that there exist no other minimal forbidden minors for P2.

This work deals with thermal-noise modeling for silicon vertical bipolar junction transistors (BJTs) and relevant integrated circuits (ICs) operating at low currents. The two-junction BJT compact model is consistently derived from the thermal-noise generalization of the Shockley semiconductor equations developed in work which treats thermal noise as the noise associated with carrier velocity fluctuations. This model describes BJT with the Itô non-linear stochastic-differential-equation (SDE) system and is suitable for large-signal large-fluctuation analysis. It is shown that thermal noise in silicon p-n-junction diode contributes to "microplasma" noise. The above model opens way for a consistent-modeling-based design/optimization of bipolar device noise performance with the help of theory of Itô's SDEs.

This paper presents a technique for disparity selection in the context of binocular pursuit. For vergence control in binocular pursuit, it is a crucial problem to find the disparity which corresponds to the target among multiple disparities generally observed in a scene. To solve the problem of the selection, we propose an approach based on histogramming the disparities obtained in the scene. Here we use an extended phase-based disparity estimation algorithm. The idea is to slice the scene using the disparity histogram so that only the target remains. The slice is chosen around a peak in the histogram using prediction of the target disparity and target location obtained by back projection. The tracking of the peak enables robustness against other, possibly dominant, objects in the scene. The approach is investigated through experiments and shown to work appropriately.

In this paper, the discrimination of ultrasonic heart (echocardiographic) images is studied by making use of some texture features, including the angular second moment, contrast, correlation and entropy which are obtained from a gray-level cooccurrence matrix. Features of these types are used as inputs to the input layer of a neural network (NN) to classify two sets of echocardiographic images-normal heart and dilated cardiomyopathy (DCM) (18 and 13 samples, respectively). The performance of the NN classifier is also compared to that of a minimum distance (MD) classifier. Implementation of our algorithm is performed on a PC-486 personal computer. Our results show that the NN produces about 94% (the confidence level setting is 0.9) and the MD produces about 84% correct classification. We notice that the NN correctly classifies all the DCM cases, namely, all the misclassified cases are of false positive. These results indicate that the method of feature-based image analysis using the NN has potential utility for computer-aided diagnosis of the DCM and other heart diseases.

This paper presents efficient algorithms for updating moving octrees with real-time performance. The first algorithm works for octrees undergoing both translation and rotation motion; it works efficiently by compacting source octrees into a smaller set of cubes (not necessarily standard octree cubes) as a precomputation step, and by using a fast, exact cube/cube intersection test between source octree cubas and target octree cubes. A parallel version of the algorithm is also described. Finally, the paper presents an efficient algorithm for the more limited case of octree translation only. Experimental results are given to show the efficiency of the algorithms in comparison to competing algorithms. In addition to being fast, the algorithms presented are also space efficient in that they can produce target octrees in the linear octree representation.