A trained sparse conjugate codebook is proposed for improving the speech quality of CELP-based coding in a noisy environment. Although CELP coding provides high quality at a low bit rate in a silent environment (creating clean speech), it cannot provide a satisfactory quality in a noisy environment because the conventional fixed codebook is designed to be suitable for clean speech. The proposed codebook consists of two sub-codebooks; each sub-codebook consists of a random component and a trained component. Each component has excitation vectors consisting of a few pulses. In the random component, pulse position and amplitude are determined randomly. Since the radom component does not depend on the speech characteristics, it handles noise better than the trained one. The trained component maintains high quality for clean speech. Since excitation vector is the sum of the two sub-excitation vectors, this codebook handles various speech conditions by selecting a sub-vector from each component. This codebook also reduces the computational complexity of a fixed codebook search and memory requirements compared with the conventional codebook. Subjective testing (absolute category rating (ACR) and degradation category rating (DCR)) indicated that this codebook improves speech quality compared with the conventional trained codebook for noisy speech. The ACR test showed that the quality of the 8 kbit/s CELP coder with this codebook is equivalent to that of the 32 kbit/s ADPCM for clean speech.

This paper describes a study to determine if a current-mode circuit is useful as an analog circuit technique for realizing submicron mixed analog-and-digital MOS LSIs. To examine this, we designed and circuit simulated a new current-mode ADC bit-block for a 3 V, 10-bit level, 20 MHz ADC with a pipeline architecture and with full current-mode approach. A new precision current-mode sample-and-hold circuit which enables operation of a bit block at a clock speed of 20 MHz was developed. Current mismatches caused by the poor output impedance of a device were also decreased by adopting a cascode configuration throughout the design. Operation with a 3 V power supply and a 20 MHz clock speed in a 3-bit A/D configuration was verified through circuit simulation using standard CMOS 0.6 µm device parameters. Gain error, mismatch of current, and linearity of the bit block with changing threshold voltage of a device were carefully examined. The bit block has a gain error of 0.2% (10-bit level), a linearity error of less than 0.1% (more than 10-bit level), and a current mismatch of DAC current sources in a bit cell of 0.2 to 0.4% (more than 8-bit level) with a 3 V power supply and 20 MHz clock speed. An 8-to 9-bit video-speed pipeline ADC can be realized without calibration. This confirms that the current-mode approach is effective.

A novel reactive congestion control scheme based on Dynamic UPC/NPC (Usage/Network Parameter Control) in ATM networks is proposed. In this scheme, policing parameters at the UPC/NPC are dynamically modified in response to the reception of RM (Resource Management) cells. In a congested state, traffic volume submitted to the network is regulated by Dynamic UPC/NPC, while providing negotiated QoS (Quality of Service) for each ATM connection. When end-stations (or edge-entities between network segments) operate according to ER-based (Explicit Rate based) behavior, a UPC/NPC function will indicate (send) an ER value toward each source end-station using backward RM cells. In this case, the policing parameter at the UPC/NPC should take the same value as the ER value. When end-stations (or edge-entities) operate according to EFCI-based (Explicit Forward Congestion Indication based) behavior, the modified policing parameter at the UPC/NPC point must be harmonized with the modified cell transmission rate at the source end-stations (or at the edge-entities). In order to improve the control performance for the long distance connections, backward RM cells will be generated by the NPC function (UPC function will be optional) at the egress of a congested network in response to the reception of EFCI marked cells (or forward RM cells) as a proxy destination end-station, and they will be sent back toward the UPC/NPC function at the ingress of the network. As a result, the proposed control scheme enables the network to recover from the congested state securely and provide the negotiated service quality, even if cooperation of (rate-based) flow control at each source end-station (and at edge-entities between network segments) is not expected.

Intelligent Network (IN) is a distributed architecture allowing telecom companies to create and to customize services. Network and services have to be integrated in the management process provided by the IN. The integration of the Intelligent Network and the TMN (Telecommunications Management Networks) can be useful to achieve the management process. Our proposal is an evolution towards a more intelligent management structure provided through the Distributed Artificial Intelligence concepts.

This letter shows that a portable visual stimulator for MEG measurements can be realized using an optical fiber bundle and a CRT display system offering high brightness and high speed raster scanning, and that MEGs with neither magnetic contamination nor jitter can be measured by the stimulator.

The message level performance of error controls in data communication on ATM network is analyzed. Three layers, "a cell"(a unit of transmission), "a block"(a unit of error controls) and "a message"(a unit of transmission of user level) are considered. The error controls treated in this paper are GBN (Go-Back-N) and FEC+GBN. The cell loss process is assumed to be the two state Markov chain considering the cell loss process in ATM networks. Numerical results show that (1) the improvement of the message forwarding delay is saturated in some environments when the interface rate becomes high, (2) FEC is efficient when the burstiness of the cell loss process is small, the message length is large and the interface rate is high.

In this study, we discuss a discrete-time cellular neural network (DTCNN) and its applications including convergence property and stability. Two theorems about the convergence condition of nonreciprocal non-uniform DTCNNs are described, which cover those of reciprocal one as a special case. Thus, it can be applied to wide classes of image processings, such as associative memories, multiple visual patterns recognition and others. Our DTCNN realized by the software simulation can largely reduce the computational time compared to the continuous-time CNN.

In this paper, we formulate and solve a discrete-time queueing problem that has two potential applications: ATM multiplexers and DQDB networks. We first consider the modeling of an ATM multiplexer. The object of the analysis is a periodic traffic stream (CBR traffic), which is one of the inputs to the multiplexer. As in previous works of the subject, we consider a memoryless background traffic input. Here, in addition to this background traffic, we take into account the influence of a high-priority traffic, which is time-correlated and requires expedited service. We analyze the influence of these two types of traffic on the statistics of the interdeparture time (jitter process) and the delay of the periodic traffic stream. We obtain their distributions in a form of z-transforms, and from these we derive closed form expressions for the average delay and the variance of the interdeparture time. Our results show that the delay and jitter are very sensitive to the burstiness of the high priority traffic arrival process. We next apply our analytical modeling to a DQDB network when some of its stations are driven by CBR sources. We can obtain interesting results concerning the influence of the physical location of a DQDB station on the jitter.

We propose a novel architecture (Switched Access Star: SAS) using an optical switch for access networks and prove its operating principle experimentally. In this architecture, the multiple optical network units (ONUs) in subscriber premises are connected to one optical subscriber unit (OSU) in a central office through an optical switch. SAS can increase the number of accommodated ONUs, the transmission line length, and the capacity per ONU. Moreover, this architecture does not need encryption or ID/passwords. SAS can reduce system cost and yield flexible transmission capacities and realize easy management and maintenance of optical transmission lines.

Distributed shared memory is an attractive option for realizing functionally distributed computing in a wide area distributed environment, because of its simplicity and flexibility in software programming. However, up till now, distributed shared memory has mainly been studied in a local environment. In a widely distributed environment, latency of communication greatly affects system performance. Moreover, bandwidth of networks available in a wide area is dramatically increasing recently. DSM architecture using high performance networks must be different from the case of low speed networks being used. In this paper, distributed shared memory models in a widely distributed environment are discussed and evaluated. First, existing distributed shared memory models are examined: They are shared virtual memory and replicated shared memory. Next, an improved replicated shared memory model, which uses internal machine memory, is proposed. In this model, we assume the existence of a seamless, multi-cast wide area network infrastructure - for example, an ATM network. A prototype of this model using multi-thread programming have been implemented on multi-CPU SPARCstations and an ATM-LAN. These DSM models are compared with SCRAMNetTM, whose mechanism is based on replicated shared memory. Results from this evaluation show the superiority of the replicated shared memory compared to shared virtual memory when the length of the network is large. While replicated shared memory using external memory is influenced by the ratio of local and global accesses, replicated shared memory using internal machine memory is suitable for a wide variety of cases. The replicated shared memory model is considered to be suitable particularly for applications which impose real time operation in a widely distributed environment, since some latency hiding techniques such as context switching or data prefetching are not effective for real time demands.

Throughout the paper, the proper operating of the self-routing principle in 2-D shuffle multistage interconnection networks (MINs) is analysed. (The notation 1-D MIN and 2-D MIN is applied for a MIN which interconnects 1-D and 2-D data, respectively.) Two different methods for self-routing in 2-D shuffle MINs are presented: (1) The application of self-routing in 1-D MINs by a switch-pattern preserving transformation of 1-D shuffle stages into 2-D shuffle stages (and vice versa) and (2) the general concept of self-routing in 2-D shuffle MINs based on self-routing with regard to each coordinate which is the original contribution of the paper. Several examples are provided which make the various problems transparent.

Threshold voltage shift in high frequency operation of 0.3µm and 0.35µm gate SOI CMOS is experimentally studied, using supply current measurement of inverter chains as test structures. The threshold voltage shift is obtained from the measurement of the leak currents in DC and high frequency condition. For a large supply voltage the electron-hole generation current becomes dominant, resulting in lowered threshold voltage, while the threshold voltage becomes higher than DC case for a low supply voltage. A reasonable relation of the threshold voltage shift and average electric field in the channel is obtained in this study. This method will be useful as a measure of "substrate current" for floating body SOI CMOS.

One of coupling coefficients appearing in the coupled power equations describing the crosstalk in an image fiber is derived based on the coupled mode theory. Cores arranged in the cross-section of the fiber differ randomly to the degree of several percent in size and consequently modes propagating along the cores differ randomly. Random fluctuations of the propagation constants of modes cause the random transfer process of power between the cores, whereas contributions of the random fluctuations of the mode coupling coefficients to the statistical process can be neglected. The coupling coefficient is described as the ratio of the power transfer ratio to the coupling length for two cores with slightly different radii characterizing the random cores. The theoretical results are in good agreement with measurement results except near cutoff.

Recursive Diagonal Torus (RDT) is a class of interconnection network for massively parallel computers with 216 nodes. In this paper, message transfer algorithms on the RDT are proposed and discussed. First, a simple one-to-one message routing algorithm called the vector routing is introduced and its practical extension called the floating vector routing is proposed. In the floating vector routing both the diameter and average distance are improved compared with the fixed vector routing. Next, broadcasting and hypercube emulation algorithm scheme on the RDT are shown. Finally, deadlock-free message routing algorithms on the RDT are discussed. By a simple modification of the e-cube routing and a small numbers of additional virtual channels, both one-to-one message transfer and broadcast can be achieved without deadlock.

Existing edge detection methods provide unsatisfactory results when contrast changes largely within an image due to non-uniform illumination. Koch et al. developed an energy function based upon the Hopfield neural network, whose coefficients were fixed by trial and error, and remain constant for the entire image, irrespective of the differences in intensity level. This paper presents an improved edge detection method for non-uniformly illuminated images. We propose that the energy function coefficients for an image with inconsistent illumination should not remain fixed, rather should vary as a second-order function of the intensity differences between pixels, and actually use a schedule of changing coefficients. The results, compared with those of existing methods, suggest a better strategy for edge detection depending upon both the dynamic range of the original image pixel values as well as their contrast.

Shortened prime codes (SPR-codes) are presented, which can maintain the fixed code weight for any arbitrary number of codewords while still preserve the same cross and auto-correlation constraints as original prime codes. The use of SPR-codes can reduce both cost and power loss of optical encoders/decoders. Tunable all-optical SPR-code encoders are also designed, which are based on rapidly tunable optical delay lines. It is shown that using this type of encoders not only can further reduce the coding power loss, but also can achieve a very cost-effective fashion.

A direct gradient descent learning algorithm of energy function in Hopfield neural networks is proposed. The gradient descent learning is not performed on usual error functions, but the Hopfield energy functions directly. We demonstrate the algorithm by testing it on an analog-to-digital conversion and an associative memory problems.

A novel technique for evaluation of charge build-up in semiconductor wafer processing such as ion implantation, plasma etching and plasma enhanced chemical vapor deposition by using the breakdown of MOS capacitors with charge collecting electrodes (antenna) is proposed. The charge build-up during high beam current ion implantation is successfully evaluated by using this technique. The breakdown sensitivity of a MOS capacitor is improved by using a small area MOS capacitor with a large area antenna electrode. To estimate charge build-up on wafers quantitatively, the best combination of gate oxide thickness, substrate type, MOS capacitor area and antenna ratio should be carefully chosen for individual charge build-up situation. The optimum structured antenna MOS capacitors which relationship between QBD and stressing current density was well characterized give us very simple and quantitative charge build-up evaluation. This technique is very simple and useful to estimate charge build-up as compared with conventional technique by suing EEPROM devices or large area MOS capacitors.

In this paper, we propose a new method of dialogue data collection which can be used to evaluate modules of a spoken dialogue system. To evaluate the module, it is necessary to use suitable data. Human-human dialogue data have not been appropriate to module evaluation, because spontaneous data usually include too much specific phenomena such as fillers, restarts, pauses, and hesitations. Human-machine dialogue data have not been appropriate to module evaluation, because the dialogue was unnatural and the available vocabularies were limited. Here, we propose 'Hybrid method' for the collection of spoken dialogue data. The merit is that, the collected data can be used as test data for the evaluation of a spoken dialogue system without any modification. In our method a human takes the role of some modules of the system and the system, also, works as the other part of the system together. For example, humans works as the speech recognition module and the dialogue management and a machine does the other part, response generation module. The collected data are good for the evaluation of the speech recognition and the dialogue management modules. The reasons are as follows. (1) Lexicon: The lexicon was composed of limited words and dependent on the task. (2) Grammar: The intention expressed by the subjects were concise and clear. (3) Topics: There were few utterances outside the task domain. The collected data can be used test data for the evaluation of a spoken dialogue system without any modification.

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.