Distributed computing over a network of workstations continues to be an illusive goal. Its main obstacle is the delay penalty due to network protocol and OS overhead. We present in this paper a low level hardware supported scheme for managing distributed shared memory (DSM), as an underlying paradigm for distributed computing. The proposed DSM is novel in that it employs a two-tier paging scheme that reduces the probability of false sharing and facilitates an efficient hardware implementation. The scheme employs a standard OS page and divides it into fixed smaller memory units called paragraphs, similar to cache lines. This scheme manages the shared data regions only, while other regions are handled by the OS in the standard manner without modification. A hardware extension of a traditional MMU, namely Distributed MMU or DMMU, is introduced to support the DSM. Shared memory coherency is maintained through a write-invalidate protocol. An analytical model is built to evaluate the system sensitivity to various parameters and to assess its performance.

A channel equalization technique on a time division duplex CDMA/TDMA system for wireless multimedia networks is investigated, and the bit error rate performance of the system is theoretically analyzed. The assumed network connects mobile terminals to a node of ATM based high speed LAN through a radio central unit. Only human interface facilities are implemented into the terminal so that users access integrated services through the node of the network. The uplink (from a mobile terminal to a radio central unit) employs a CDMA scheme to transmit human interface signals and the downlink employs a TDMA scheme to transmit display interface signals. Both the CDMA and the TDMA signals occupy the same frequency band. To mitigate bit error rate degradation due to fading, the radio central unit estimates the impulse response of the channel from the received CDMA signals and subtracts the replica signal to cancel the major intersymbol interference (ISI) component. Numerical results using the Nakagami-m fading model and recent propagation measurements show that the proposed TPC technique compensates the fading attenuation and the proposed CEQ cancels the major ISI component. The bit error rate performance of the downlink with the proposed CEQ is superior to that with the DFE by 12dB of the symbol SNR at the BER=10-6 over a specular channel, and the system with the proposed CEQ achieves a BER=10-6 at the symbol SNR=12dB. Furthermore, the channel equalizer is implemented without increases in complexity of the terminal because all the processing on the equalization is carried out only in the radio central unit.

This paper presents an improved neural network for channel assignment problems in cellular mobile communication systems in the new co-channel interference model. Sengoku et al. first proposed the neural network for the same problem, which can find solutions only in small size cellular systems with up to 40 cells in our simulations. For the practical use in the next generation's cellular systems, the performance of our improved neural network is verified by large size cellular systems with up to 500 cells. The newly defined energy function and the motion equation with two heuristics in our neural network achieve the goal of finding optimum or near-optimum solutions in a nearly constant time.

The back-propagation algorithm has been applied to many fields, and has shown large capability of neural networks. Many people use the back-propagation algorithm together with a momentum term to accelerate its convergence. However, in spite of the importance for theoretical studies, theoretical background of a momentum term has been unknown so far. First, this paper explains clearly the theoretical origin of a momentum term in the back-propagation algorithm for both a batch mode learning and a pattern-by-pattern learning. We will prove that the back-propagation algorithm having a momentum term can be derived through the following two assumptions: 1) The cost function is Enαn-µEµ, where Eµ is the summation of squared error at the output layer at the µth learning time and a is the momentum coefficient. 2) The latest weights are assumed in calculating the cost function En. Next, we derive a simple relationship between momentum, learning rate, and learning speed and then further discussion is made with computer simulation.

A subharmonic injection-locked oscillator (ILO) MMIC chain is proposed for the local oscillators and synthesizers used at millimeter-wave frequencies. A fabricated, primary 11-GHz-band injection-locked oscillator MMIC for the first stage ILO in the ILO-chain MMIC, achieves a wide subharmonic-injection-locking range at the subharmonic factors, 1/n (n=1, 2, 3, ), of 1/1, 1/2 and 1/3. The ILO MMIC abilities for synthesizer applications were confirmed with an injection-locking time of only 100-200 nsec, which is less than 1/100 that of PLL oscillators, and also with free-running oscillation performance and a wide injection locking range within a temperature range of -30 and 80.

This paper discusses advanced wireless communication technologies for achieving future high-speed mobile radios. Mainly, five technical fields are considered, that is, multi-level modulation for transmitting high-capacity information signal, advanced adaptive wireless system flexibly changing modulation level, symbol rate and traffic according to fading conditions, adaptive multicarrier system transmitting multimedia signals by changing the number of carrier according to the capacity of the signals, new CDMA techniques for mapping different bit rate services onto the same allocated bandwidth at the same time, and optical-linked microcellular communication system with millimeter wave air interface.

This paper proposes an unsupervised speaker adaptation method using an all-phoneme ergodic Hidden Markov Network" that combines allophonic (context-dependent phone) acoustic models with stochastic language constraints. Hidden Markov Network (HMnet) for allophone modeling and allophonic bigram probabilities derived from a large text database are combined to yield a single large ergodic HMM which represents arbitrary speech signals in a particular language so that the model parameters can be re-estimated using text-unknown speech samples with the Baum-Welch algorithm. When combined with the Vector Field Smoothing (VFS) technique, unsupervised speaker adaptation can be effectively performed. This method experimentally gave better performances compared with our previous unsupervised adaptation method which used conventional phonetic HMMs and phoneme bigram probabilities especially when the amount of training data was small.

In case of object recognition using 3-D configuration data, the scale and poses of the object are important factors. If they are not known, we can not compare the object with the models in the database. Hence we propose a strategy for object recognition independently of its scale and poses, which is based on Hopfield neural network. And we also propose a strategy for estimation of the camera motion to reconstruct 3-D configuration of the object. In this strategy, the camera motion is estimated only with the sequential images taken by a moving camera. Consequently, the 3-D configuration of the object is reconstructed only with the sequential images. And we adopt the multiple regression analysis for estimation of the camera motion parameters so as to reduce the errors of them.

This paper presents a system for recognition of on-line cursive Hangul (Korean characters) by means of DP matching of structural information. The penalty function has the following special features. In order to prevent short spurious strokes from causing large penalties, an input stroke is weighted by its length relative to other input strokes. In order to make use of pen-up and pen-down information, a penalty is incurred when 2 strokes of differing type (i.e. pen-up with pen-down) are matched. Finally, to reduce the chance of obtaining a suboptimal solution which can result from using the greedy algorithm in DP matching, we look-ahead an extra match. In a computer simulation we obtained a recognition rate of 92% for partially cursive characters and 89% for fully cursive characters. Furthermore, for both cases combined the correct character appears 98% of the time in the top 10 candidates. Thus we confirmed that the proposed algorithm is effective in recognizing cursive Hangul.

An error correction/detection decoding scheme of binary Hamming codes is proposed. Error correction is performed by algebraic decoding and then error detection is performed by simple likelihood ratio testing. The proposed scheme reduces the probability of undetected decoding error in comparison with conventional error correction scheme and increases throughjput in comparison with conventional error detection scheme.

An optimum design scheme for logical network topologies on a Flexible Multi-QoS Logical ATM Network, named Full-Net, is proposed. Full-Net offers high-quality Virtual-Path (VP) networks and controls end-to-end QoS only at the VP-network's access points. To develop the optimum network topology for multimedia traffic in a single ATM network, a logically con figured Virtual Channel Handler (VCH) interconnection network is associated with each QoS class. Many logical networks can be mapped at the same time on the same network, because mapping is independent of the network's physical implementation. To achieve an optimum design scheme for logical networks, the number of disjoint routes is introduced as the parameter used to optimize logical network topology. The number of disjoint routes is chosen so as to maximize total network efficiency. The optimum number of disjoint routes depends on the required QoS, VC-traffic characteristics, and traffic demand. By choosing the relevant cost characteristics, the network operator can easily maximize network efficiency and provide customers with the QoS they request at minimum cost. The proposed optimum multi-QoS network design scheme on a Full-Net architecture is an efficient solution to implementing multi-QoS control in an ATM network.

The stability of a high-resistivity load SRAM cell using thin-film SOI MOSFET's was investigated as compared with bulk-Si MOSFET's. In SOI MOSFET's back-gate-bias effect was suppressed by indirect application of back-gate-bias to the channel region through the thick buried oxide. The Vt shifts were reduced to be 10% and 14% of that in bulk-Si MOSFET's in partially and fully depleted devices, respectively. The reduction of back-gate-bias effect provides improvement of "high" output voltage and gain in the enhancement-enhancement (EE) inverter in a high-resistivity load SRAM cell, thereby offering improved cell stability. It was demonstrated by using partially depleted SOI SRAM cells that non-destructive reading was obtained even at a low drain voltage of 1.4 V without gate-potential boost, which was much smaller than the operation limit in a bulk Si SRAM with the same patterns and dimensions used as a reference. This implies that SOI devices can also offer low-voltage operation even in TFT-load cells used in up-to-date high-density SRAM's. These results suggest that thin-film SOI MOSFET's have a superior potential of low-voltage operation expected for further scaled devices and/or for portable systems in a forthcoming multimedia era.

Multimedia communication services are being made available with the advent of broadband optical fiber networks. As many different services will be accommodated in such networks, network survivability has been recognized to be a crucial concern. In this paper, we propose a new restoration algorithm for ATM networks providing multimedia services. Our proposed restoration algorithm adopts the message bundling scheme of the Multi-Destination Flooding (MDF) algorithm which was previously proposed for STM-based networks to handle catastrophic failures such as multiple link and node failures. Virtual Paths (VP) with the same communication speed are bundled and Operation Administration and Maintenance (OAM) cells are used for communication of restoration messages. In addition, the following modifications are made on the original MDF to improve restoration performance. The pre-cancellation scheme is adopted to arbitrate reservation contention to realize high restoration ratio. The dual queue scheme is applied to avoid congestion of restoration messages. Moreover, the connection control scheme for VPI connections is proposed to prevent alternative routes from being misconnected. This paper describes the design concept of our restoration algorithm, processes in each restoration phase, and the performance evaluation by computer simulation.

This paper discussed a distributed processing architecture for a virtual space teleconferencing system. Virtual space teleconferencing is a promising application field for networked virtual environments. People in different places will be able to meet each other in a virtual teleconferencing room and proceed with various cooperative tasks. When such a system creates a realistic virtual environment, it can be referred to as a "Teleconferencing with realistic sensations" system. Further more, as the conference environment can be shared by a number of users, it is possible to perform various kinds of cooperative work using the system. In this paper, the architecture for networked multi-user virtual space systems are classified, and then a case study is described for building a proposed teleconferencing system. The system reproduces a 3D image of each conference participant in a virtual meeting room. Compared with the former system, the new system can deal with more than three participants at the same time and can connect them through commercial telephone lines. Based on the virtual world database management structure, the system was classified as a central server system. However, a central server architecture limits the number of conference sites. We confirmed that the system can serve up to 14 sites using multi-modal interaction without significant latency in operation from summational experiments. Then, introducing some assumptions to the results, we have proposed processing model of the system. The results of model could describe the experimental results and we could indicate roughy estimated system capacity to realize a requaied system performance.

Recently, through high speed computer networks, multifarious information such as text, moving and still images, video, voice and control data is available. There is a natural demand to store such multimedia data in databases to facilitate their reuse in a wide variety of applications. Therefore, important research issues pertain to the investigation of appropriate database systems in multimedia computer network environments. In this paper, we first discuss the required technologies for multimedia information systems. Then we look at many multimedia information services through computer networks, and consider the importance of storing and effectively reusing such available multimedia data. To facilitate developing databases for use in these environments, we discuss the evolution of the notion of database systems. Finally, we demonstrate, as practical examples of such database systems, two prototype systems that we are currently implementing, i.e., the campus wide news on demand system and the ASN. 1 database system.

A novel wireless multimedia network employing a time division duplex CDMA/TDMA scheme is proposed. The network connects mobile multimedia terminals to an ATM based LAN through a radio central unit, and provides both uplink and downlink with unbalanced data rates in the same frequency band. The uplink (from a mobile terminal to a radio central unit) employs a CDMA scheme to transmit low speed human interface signals (-2.4kbit/s), and the downlink employs a TDMA scheme to transmit high speed video signals (-24Mbit/s). The data rates of both links are independent from that of the LAN. The uplink also employs a RAKE receiver and a forward error correction (FEC) scheme using a BCH code in order to reduce bit errors caused by multipath fading. To mitigate channel degradation caused by the near-far problem and multipath fading, a transmission power control (TPC) method for both links and a channel equalizer (CEQ) for the downlink are proposed. The control signals for the TPC and the CEQ are estimated from the impulse response of the channel which is extracted as the output of the matched filters in the CDMA receiver. Theoretical analyses are performed to evaluate the bit error rate (BER) characteristics of the proposed network. The BER performance is derived for a general multipath fading condition modeled by the Nakagami-m distribution and a typical delay profile. Numerical calculation using recent propagation measurements shows the bit error rates of both uplink and downlink to be less than 10-6 when both of the TPC and the CEQ are employed if there are some specular components in the received signals. This excellent performance can cut a way to realize a mobile multimedia terminal for customer premises. Furthermore, the configuration of the mobile terminal is quite simple even if the high speed TDMA signals are received over a multipath fading channel.

Currently there is considerable world-wide speculation regarding the introduction of optical fiber cable into access networks, because optical fiber has a big potential for providing attractive multimedia services. Since optical fiber cable can provide a variety of grade of services, high-reliability of cable networks would be required compared with the conventional copper cable networks. To develop multimedia telecommunication networks as an infrastructure, it is urgent to clarify the highly reliable optical fiber cable network architecture. Since cable network architecture deeply depends on regional conditions such as demand, area size, duct layer networks (consisting of ducts, manholes, tunnels, feeder points etc.), it is necessary to develop a cable network designing tool with user-friendly interfaces for efficiently evaluating cable network architectures. This paper firstly proposes the heuristic algorithms enhanced by the disjoint-shortest-path and the depth-first-search methods that would be applicable for real access networks. Secondly, the design method of highly reliable optical fiber cable network based on the heuristic algorithms in terms of network cost and unavailability caused by cable breakdown is proposed. It can design the combination of star- and loop-shaped (where two diversified routes exist between a feeder point and central office) cable network. Furthermore, comparison with the conventional design method which simply applies star- or loop-shaped cable network is done in terms of economy and reliability on real access networks in the Tokyo metropolitan area. It is concluded that the proposed method can reduce the network cost further and realize a short unavailability value compared with the conventional method.

The dynamics of an artificial neural network derived from a biological system, and its two applications to engineering problems are examined. The model has a multi-layer structure simulating the primary and secondary components in the olfactory system. The basic element in each layer is an oscillator which simulates the interactions between excitatory and inhibitory local neuron populations. Chaotic dynamics emerges from interactions within and between the layers, which are connected to each other by feedforward and feedback lines with distributed delays. A set of electroencephalogram (EEG) obtained from mammalian olfactory system yields aperiodic oscillation with 1/f characteristics in its FFT power spectrum. The EEG also reveals abrupt state transitions between a basal and an activated state. The activated state with each inhalation consists of a burst of oscillation at a common time-varying instantaneous frequency that is spatially amplitude-modulated (AM). The spatial pattern of the activated state seems to represent the class of the input ot the system, which simulates the input from sensory receptors. The KIII model of the olfactory system yields sustained aperiodic oscillation with "1/f" spectrum by adjustment of its parameters. Input in the form of a spatially distributed step funciton induces a state transition to an activated state. This property gives the model its utility in pattern classification. Four different methods (SD, RMS, PCA and FFT) were applied to extract AM patterns of the common output wave forms of the model. The pattern classification capability of the model was evaluated, and synchronization of the output wave form was shown to be crucial in PCA and FFT methods. This synchronization has also been suggested to have an important role in biological systems related to the information extraction by spatiotemporal integration of the output of a transmitting area of cortex by a receiving area.

This paper presents a new structure for noise and echo cancelers based on a combined fast abaptive algorithm. The main purpose of the new structure is to detect both the double-talk and the unknown path change. This goal is accomplished by using two adaptive filters. A main adaptive filter Fn, adjusted only in the non-double-talk period by the normalized LMS algorithm, is used for providing the canceler output. An auxiliary adaptive filter Ff, adjusted by the fast RLS algorithm, is used for detecting the double-talk and obtaining a near optimum tap-weight vector for Fn in the initialization period and whenever the unknown path has a sudden or fast change. The proposed structure is examined through computer simulation on a noise cancellation problem. Good cancellation performance and stable operation are obtained when signal is a speech corrupted by a white noise, a colored noise and another speech signal. Simulation results also show that the proposed structure is capable of distinguishing the near-end signal from the noise path change and quickly tracking this change.

We investigate the multipacket message transmissions and variable length message transmissions in slotted ALOHA systems with capture effect. First, we propose an approach that the transmission power level is controlled probabilistically depending on message length for multipacket messages. We consider the multipacket messages model with capture. We derive explicit equations of the effective channel utilization of the model. It is demonstrated that if we increase the numbar of power levels, we can get more effective channel utilization of the system. Secondly, we propose how to assign the slot size and show that the effective utilization of the channel is improved for variable length messages using the approach proposed for multipacket messages. Channel design issue about length of the slot depending on the number of power levels used for transmission is discussed. Thirdly, we propose the multiple messages per slot model with capture. The analytical results show that the multiple messages per slot model can achieve the highest channel utilization among the models discussed in this paper.