We propose a new CALL (Computer-Assisted Language Learning) system for non-native learners of Japanese using speech recognition methods. The aim of the system is to help them develop natural pronunciation by automatically detecting their pronunciation errors and then providing effective feedback instruction. An automatic scoring method based on HMM log-likelihood is used to assess their pronunciation. Native speakers' scores are normalized by the mean and standard deviation for each phoneme and are used as threshold values to detect pronunciation errors. Unlike previous CALL systems, we not only detect pronunciation errors but also generate appropriate feedback to improve them. Especially for the feedback of consonants, we propose a novel method based on the classification of the place and manner of articulation. The effectiveness of our system is demonstrated with preliminary trials by several non-native speakers.

A 50-Gbit/s demultiplexer IC module that uses 0.1-µm InAlAs/InGaAs/InP HEMTs is reported. The maximum error-free operation bit-rate of a fabricated module is 50 Gbit/s, and a wide phase margin of 170 degrees is obtained at 43 Gbit/s. 50-Gbit/s demultiplexing is the fastest performance of all packaged demultiplexer ICs yet reported.

To enhance fault tolerance ability of the feedforward neural networks (NNs for short) implemented in hardware, we discuss the learning algorithm that converges without adding extra neurons and a large amount of extra learning time and cycles. Our algorithm modified from the standard backpropagation algorithm (SBPA for short) limits synaptic weights of neurons in range during learning phase. The upper and lower bounds of the weights are calculated according to the average and standard deviation of them. Then our algorithm reupdates any weight beyond the calculated range to the upper or lower bound. Since the above enables us to decrease the standard deviation of the weights, it is useful in enhancing fault tolerance. We apply NNs trained with other algorithms and our one to a character recognition problem. It is shown that our one is superior to other ones in reliability, extra learning time and/or extra learning cycles. Besides we clarify that our algorithm never degrades the generalization ability of NNs although it coerces the weights within the calculated range.

In this paper, the authors derived the distributions of the probability of detection and of false alarm in function of the decision threshold. An Optimized Threshold Decision (OTD) algorithm was proposed to decide the optimal threshold for reaching the best system performance in a given known channel noise. By applying this OTD algorithm, the multiple access capacity can thus be maximized.

In this paper, we consider the coherent radar detection of rapid fluctuating signals in the Gaussian noise. This problem has been previously solved by employing the GLR technique, but we use the ALR to improve the detection performance. So, after deriving an approximate ALR detector, we compare the new detector with the GLR and Square-law detectors and we show its superiority.

Petri net is a mathematical model for concurrent systems. Petri net is known to have less modeling power than that of Turing machine. Lack of zero testing ability is the main reason of this fact. Indeed, every extended Petri net model that can perform zero testing is equivalent to Turing machine. Time Petri net is one of the models with ability of zero testing. It is of theoretical interest what structure enables zero testing. This paper shows that time asymmetric choice net can simulate register machines. The result implies reachability problem of this subclass of time Petri net is undecidable.

This paper presents the bit error rate (BER) upper bounds for trellis coded asymmetric 8PSK (TC-A8PSK) system using the Ka-band satellite in the rain fading environment. The probability density function (PDF) for the rain fading random variable can be theoretically derived by assuming that the rain attenuation can be approximated to a log-normal distribution and the rain fading parameters are calculated by using the rain precipitation data from the Crane global model. Furthermore, we analyze the BER upper bounds of TC-A8PSK system according to the number of states in the trellis diagram and the availability of channel state information (CSI). In the past, Divsalar and Simon has analyzed the BER upper bounds of 2-state TCM system in Rician fading channels; however, this paper is the first to analyze the BER upper bounds of TCM system in the rain fading channels. Finally, we summarize the dominant six factors which are closely related to the BER upper bounds of TC-A8PSK satellite system in the rain fading channel as follows: (1) frequency band, (2) rain intensity, (3) elevation angle, (4) bit energy to noise ratio, (5) asymmetric angle, and (6) availability of CSI.

The Internet is a heterogeneous and rapidly evolving environment. New information services have been added, modified, and diversified, while legacy services have been abandoned. Users' requirements for services are also changing very quickly. It is becoming very difficult to find required information services from extremely huge amount of information in the Internet. It is necessary to design information service system, which can meet user's heterogeneous requirements for services, adaptability of services to cope with ever-changing situation and to provide assurance. Faded information field architecture and Push/Pull Mobile Agent (MA) technologies are proposed as basic components of autonomous information service system to achieve these goals. The system is proposed from users' and providers' point of view, therefore we present a technique for balancing users' and providers' communication cost based on `through rate' for faded information field realization. On the basis of this technique, autonomous information allocation algorithm is proposed, in which each node may determine information amount based on push/pull cost. The effectiveness of the proposed algorithm is shown through simulation.

To provide cellular data services with differentiated QoS, a shared resource scheme, based on the optimization theory and LaGrange λ-calculus was developed. This scheme can generate a fair schedule for a diverse mix of traffic with diverse QoS requirements in a limited radio spectrum. We define the acceptance indication, AI, as the QoS measurement for the shared resource scheme. The experimental results show that this approach outperforms other existing schemes.

This letter proposes a modified synchronous acquisition method using code-orthogonalizing filters (COFs) in an asynchronous direct sequence (DS) / CDMA. Improvements on the average acquisition time for several conditions are shown.

In a typical indoor environment such as in a building, delay spread tends to be small, which causes frequency non-selective fading. Therefore resolvable paths at the RAKE receiver can not be obtained, and effective path diversity can not be achieved. This paper proposes an artificial path diversity system in which one or multiple sectors at the base station are pre-selected according to the channel conditions for transmitting data. Each sector's signal is delayed by several chips to create artificial paths which can then be combined by using a RAKE receiver at the mobile station creating a diversity effect for an indoor environment. Moreover, only pre-selected sector antennas transmit signals to reduce inefficient signal usage in the sectors whose paths are blocked by using all sectors, therefore the transmission power is used efficiently at the base station. As a result of sector selection, the better BER performance and the reduction of interference signals between different channels can be achieved by means of sector selection. The performance of the proposed system is analyzed and demonstrated by computer simulation in a Rayleigh and log-normal fading indoor environment.

We propose a unique word detection scheme having the correction of the phase ambiguity that remains in the output of a carrier recovery circuit for QPSK systems. The performance of the proposed scheme is analyzed by obtaining the probability generating function (PGF) for the output of the detector. The PGF is obtained from the z-transform of the probability density function (PDF) for each matched symbol with the reference symbol and by using the characteristics at this time that the random variables for matched symbols are independent and identically distributed. Then the total PGF for a unique word is obtained from powering the PGF for the individual symbol. From the total PGF, the length of a UW and the threshold value for detection can be obtained with a detection probability defined in a required system specification.

The Real-time VLBI Correlator (RVC) is a new type processor for the Very-Long-Baseline Interferometry (VLBI). This correlator was primarily designed for supporting the VLBI Space Observatory Programme (VSOP). Two particular techniques, the fringe rotator after correlation and the lag-time extension technique, are newly developed for the RVC. The correlation circuit size of VLBI correlator is reduced to half by introducing the new fringe rotator, and it makes possible to realize a large delay window being essential in finding a cross correlation in real-time. The delay window can be changed flexibly with the lag-time extension technique, and its technique is useful to detect the fringe peak in a VSOP observation. The new correlator was installed at the Usuda Deep Space Center in Japan, and is used in VSOP and other domestic VLBI observations. In this paper, the key features of the Real-time VLBI Correlator (RVC) focusing on these advanced techniques are presented, and the results of its performance test are shown.

Numerous scientific and engineering fields extensively utilize optimization techniques for finding appropriate parameter values of models. Various optimization methods are available for practical use. The optimization algorithms are classified primarily due to the rates of convergence. Unfortunately, it is often the case in practice that the particular optimization method with specified convergence rates performs substantially differently on diverse optimization tasks. Theoretical classification of convergence rates then lacks its relevance in the context of the practical optimization. It is therefore desirable to formulate a novel classification framework relevant to the theoretical concept of convergence rates as well as to the practical optimization. This article introduces such classification framework. The proposed classification framework enables specification of optimization techniques and optimization tasks. It also underlies its inherent relationship to the convergence rates. Novel classification framework is applied to categorizing the tasks of optimizing polynomials and the problem of training multilayer perceptron neural networks.

If the direction of arrival (DOA) of a source is perturbed due to some reasons in a statistical way as in the environment of wireless mobile communications, a new model appropriate for such environment should be used instead of the point source model. In this paper, an angle-perturbed source model is proposed and an estimation method based on the eigen-decomposition technique is investigated under the model. The asymptotic distribution of the estimation errors is obtained to see the statistical properties.

First order line seach optimization techniques gained essential practical importance over second order optimization techniques due to their computational simplicity and low memory requirements. The computational excess of second order methods becomes unbearable for large optimization tasks. The only applicable optimization techniques in such cases are variations of first order approaches. This article presents one such variation of first order line search optimization technique. The presented algorithm has substantially simplified a line search subproblem into a single step calculation of the appropriate value of step length. This remarkably simplifies the implementation and computational complexity of the line search subproblem and yet does not harm the stability of the method. The algorithm is theoretically proven convergent, with superlinear convergence rates, and exactly classified within the formerly proposed classification framework for first order optimization. Performance of the proposed algorithm is practically evaluated on five data sets and compared to the relevant standard first order optimization technique. The results indicate superior performance of the presented algorithm over the standard first order method.

In this paper, we study macro/micro diversity techniques for code acquisition of a direct-sequence spread-spectrum signal in an indoor packet communication system. In the system discussed, the base station has several radio ports each with a cluster of antennas, and the terminal also has multiple antennas. The performance in the uplink of this system is analyzed under Lognormal shadowing and flat Rayleigh fading. The numerical results show great performance improvements by proposed diversity techniques. In addition, it is clarified that the mean acquisition time, which is often used as the measure of performance, is not suitable for packet radio systems as it underestimates the necessary preamble length for initial code acquisition.

This paper proposes aspiration controls which restrains aspiration branches and keeps the original tabu-based searching by considering past and/or (predicted) future searching profiles. For implementation of the aspiration control we employ not only the short-term and long-term memory but also future memory which is first introduced in this paper as a new concept in the tabu search field. The tabu search with the aspiration control is also parallelized. Moreover two types of parallel cooperative searching scheme are proposed. Through computational experiment, we observe efficiency of our approach comparing to the traditional ones. Especially, we find that cooperative searching has possibility to improve the solution quality very well.

With the rising innovative antigens (such as intruders and viruses) through Internet, reliable security mechanisms are required to perceptively detect and put them down. However, defense techniques of the current host system over Internet may not properly analyze Internet antigens, because trends of attacks are unexpectedly shifted. In this paper, we introduce an Antibody Layer that mediates proper security services based on the biological mechanism to rapidly disclose and remove innovative antigens. The proposed Antibody Layer also employs a new topology called antibody cooperation protocol to support real-time security QoS for one host as well as host alliance.

Recently we have studied the performance of the VS/VD behavior for two binary feedback congestion control schemes, FECN and BECN, when they are applied to multicast ABR services. It is shown that the VS/VD schemes alleviate the problems of consolidation noise and consolidation delay, which are two major design issues in multicast congestion control. To complement, we further explore the VS/VD behavior for the explicit rate feedback control scheme in this letter. Through an efficient rate estimation method, it is again observed that VS/VD behavior performs better than the traditional end-to-end schemes.