In this paper, we describe an adaptive integration method for an audio-visual speech recognition system that uses not only the speaker's audio speech signal but visual speech signals like lip images. Human beings communicate with each other by integrating multiple types of sensory information such as hearing and vision. Such integration can be applied to automatic speech recognition, too. In the integration of audio and visual speech features for speech recognition, there are two important issues, i.e., (1) a model that represents the synchronous and asynchronous characteristics between audio and visual features, and makes the best use of a whole database that includes uni-modal, audio only, or visual only data as well as audio-visual data, and (2) the adaptive estimation of reliability weights for the audio and visual information. This paper mainly investigates two issues and proposes a novel method to effectively integrate audio and visual information in an audio-visual Automatic Speech Recognition (ASR) system. First, as the model that integrates audio-visual speech information, we apply a product of hidden Markov models (product HMM), the product of an audio HMM and a visual HMM. We newly propose a method that re-estimates the product HMM using audio-visual synchronous speech data so as to train the synchronicity of the audio-visual information, while the original product HMM assumes independence from audio-visual features. Second, for the optimal audio-visual information reliability weight estimation, we propose a Gaussian mixture model (GMM) based-MCE-GPD (minimum classification error and generalized probabilistic descent) algorithm, which enables reductions in the amount of adaptation data and amount of computations required for the GMM estimation. Evaluation experiments show that the proposed audio-visual speech recognition system improves the recognition accuracy over conventional ones even if the audio signals are clean.

This paper presents a new hierarchical scheduling method for a large-scale manufacturing system based on the hybrid Petri-net model, which consists of CPN (Continuous Petri Net) and TPN (Timed Petri Net). The study focuses on an automobile production system, a typical large-scale manufacturing system. At a high level, CPN is used to represent continuous flow in the production process of an entire system, and LP (Linear Programming) is applied to find the optimal flow. At a low level, TPN is used to represent the manufacturing environment of each sub-production line in a decentralized manner, and the MCT algorithm is applied to find feasible semi-optimal process sequences for each sub-production line. Our proposed scheduling method can schedule macroscopically the flow of an entire system while considering microscopically any physical constraints that arise on an actual shop floor.

Although TCP is widely used in the Internet, new specifications are still proposed and implemented. In the circumstance above, it is highly possible that some errors are detected on the communication between new and old implementations. Several test tools were developed so far. However, they do not have enough functions to allow test operators to modify test sequences suitable for their test purposes. We have developed a TCP tester which generates test sequence using test scenario. The tester performs exceptional TCP protocol behavior only when the condition specified in the test scenario is satisfied. Otherwise, it performs normal TCP behavior. The tester is implemented by modifying TCP module of NetBSD with SACK code developed by Pittsburgh Supercomputing Center. We have also evaluated implementations of congestion control and SACK algorithms using the tester.

The characteristics of 155 Mbps optical propagation through atmosphere at the wavelengths of 0.78 and 1.55 m over a 72 m range were measured. Two propagation characteristics were compared to examine the wavelength dependence on the rain rate and the aperture size of a receiver. An enhanced intensity exceeding the signal level measured in the absence of rain was observed at both wavelengths during 7.2 mm/h rain. The resultant bit error rate is shown to be dominated mainly by the rain-induced scintillation rather than the attenuation.

One of the most important technical problems in reliable multicast protocol is reducing redundant feedback information, e.g. NAKs, to avoid feedback implosion. A number of feedback suppression mechanisms have been proposed to deal with this problem. In the MBONE, which is a virtual multicast network and makes multi-point communication across the Internet feasible, the source link, the links directly connected to or very close to the source, is reported to contribute high percentile packet loss. When a well-known NAK suppression mechanism is applied, in the case of the source link loss, all receivers suffer the same packet loss and NAK suppression mechanism does not work effectively. In this paper we propose a Reliable Multicast Protocol Applying Local FEC, called Local FEC, where the source link loss is recovered with FEC applied locally only to the source link. To investigate performance of our proposed protocol, it is compared with a popular reliable multicast protocol with NAK suppression mechanism. Our performance analysis results with mathematical analysis and computer simulation show that our proposed protocol outperforms the NAK suppression protocol from the viewpoint of scalability and wasted bandwidth.

We newly propose an optical coherence domain reflectometry for optical subscriber networks with measurement range enhancement. This reflectometry is based on our own technique to synthesize an optical coherence function. An optical switch after a light source generates optical pulses, which select the measuring region, in which one coherence peak is scanned with high spatial resolution. An optical fiber loop delay line including a frequency shifter is placed in a reference-path of the interferometer. In this method, the measuring region could be easily changed by the hetelodyne intermediate frequency selected at the electronic band pass filter. In the basic experiments, the reflections at 5 km distance are measured with a spatial resolution of 8 cm, and the change of the measuring region is successfully demonstrated.

In the past, enhancement techniques for the end-to-end quality of a networked application were studied by looking at each individual layer. Examples of such techniques include the error resilience/concealment methods in the application layer, the FEC/ARQ in the transport layer, and the Quality of Service (QoS) techniques in the network layers. However, an integrated approach that would look across all related layers had yet to be investigated. This paper proposes an approach that combines priority-aware video packetization, adaptively used layered FEC, and QoS controlled networks such as IntServ and DiffServ in order to provide an efficient end-to-end quality in layered streaming video. The combination is more efficient in terms of a simple network price mechanism, that is, in getting the best end-to-end quality under a given total cost constraint. Our proposed approach in DiffServ with video packet (VP) data-splitting and layered FEC guarantees minimal service quality in a scalable and cost effective manner without introducing resource reservation. For video, we also propose performance metrics such as corrupted/frozen frame ratio (CFR, FFR). These provide objective metrics like PSNR as well as a measurement for subjective perceptions. Our approach, which combines related layers such as video coding, transport, and network, has yielded results that have proven to be more cost-effective and practical than the supporting network QoS.

The low phase noise frequency source to be used for measurements and so on realizes by oscillator having highly output signal power against output noise power. SAW devices can be used with high power than BAW devices. So we examine on configuration of SAW oscillator circuits with the power gain. In this paper we shall discuss a configuration of oscillator circuit to obtain an extremely low phase noise and an oscillator operating at a non-reactive frequency of SAW resonator.

As pseudorandom number generators for Monte Carlo simulations, inversive linear congruential generators (ICG) have some advantages compared with traditional linear congruential generators. It has been shown that a sequence generated by an ICG has a low discrepancy even if the length of the sequence is far shorter than its period. In this paper, we formulate fractional linear congruential generators (FCG), a generalized concept of the inversive linear congruential generators. It is shown that the sequence generated by an FCG is a geometrical shift of a sequence from an ICG and satisfies the same upper bounds of discrepancy. As an application of the general formulation, we show that under certain condition, "Leap-Frog technique," a way of splitting a random number sequence to parallel sequences, can be applied to the ICG or FCG with no extra cost on discrepancy.

The aim of this paper is to present a new cost estimation technique to synthesis hardware from high level circuit description. The scheduling and allocation processes are performed in alternative manner, while using realistic cost measurements models that account for Functional Unit (FU), registers, and multiplexers. This is an improvement over previous works, were most of them use very simple cost models that primarily focus on FU resources alone. These latest, however, are not accurate enough to allow effective design space exploration since the effects of storage and interconnect resources can indeed dominates the cost function. We tested our technique on several high-level synthesis benchmarks. The results indicate that the tool can generate near-optimal bus-based and multiplexer-based architectural models with lower number of registers and buses, while presenting high throughput.

This paper deals with a method for designing the driving circuit of an electroluminescent (EL) element whose power consumption is lower and the deviation of output voltage is smaller even when the EL element is replaced with another of a different shape. In this driving circuit, an AC voltage raised by a step-up transformer is supplied to the EL element. The oscillation conditions in the orthodox driving circuits were theoretically analyzed, and a new driving circuit which incorporates these characteristics is proposed. A new prototype driving circuit taking the resonance characteristics between the capacitance of the element and the inductance of the transformer into consideration was made. In the experiment, an inorganic AC EL cable-type element was used as the load of the driving circuit as its impedance can be easily adjusted by changing its length. In comparison with orthodox circuits, the power consumption was lower and the changes in the output voltage were smaller in the new prototype circuit even when the changes in the impedance were large.

This paper introduces a way to realize high-pass, band-stop and all-pass transfer functions using two-integrator loop structure consisting of loss-less and lossy integrators. The basic circuit configuration is constructed with five Operational Transconductance Amplifiers (OTAs) and two grounded capacitors. It is shown that the circuit can realize their circuit transfer functions by choosing the input terminals, and that the circuit parameters can also be independently set by the transconductance gains with the proportional block. Although the basic circuit configuration has been known, it seems that the feature for realizing the high-pass, the band-stop and the all-pass transfer functions makes the structure more attractive and useful. An example is given together with simulated results by PSPICE.

In 1999, Araki et al. proposed a convertible limited verifier signature scheme. In this letter, we propose a universal forgery attack on their scheme. We show that any one can forge a valid signature of a user UA on an arbitrary message.

The use of fluid Generalized Processor Sharing (GPS) algorithm for integrated service networks has received much attention since early 1990's because of its desirable properties in terms of delay bound and service fairness. Many Packet Fair Queuing (PFQ) algorithms have been developed to approximate GPS. However, owing to the implementation complexity, it is difficult to support a large number of sessions with diverse service rates while maintaining the GPS properties. The grouping architecture has been proposed to dramatically reduce the implementation complexity. However, the grouping architecture can only support a fixed number of service rates, thus causing the problems of granularity, bandwidth fairness, utilization, and immunity of flows. In this paper, we propose a new implementation approach called dual-rate grouping, which can significantly alleviate the above problems. Compared with the grouping architecture, the proposed approach possesses better performance in terms of approximating per session-based PFQ algorithms without increasing the implementation complexity.

This paper compares by computer simulation the achievable throughput performance employing fast packet scheduling algorithms focusing on the throughput of each user in High Speed Downlink Packet Access (HSDPA). Three packet scheduling algorithms are employed: the Maximum carrier-to-interference power ratio (CIR), Proportional Fairness (PF), and Round Robin (RR) methods. The simulation results elucidate that although the Maximum CIR method achieves an aggregated user throughput within a cell higher than that using the PF and RR methods, the PF method is advantageous because it enhances the user throughput for a large number of users with a lower received signal-to-interference power ratio (SIR), who are located outside the normalized distance of 0.6-0.7 from a cell site (this corresponds to the area probability of 50-60% within the cell) compared to the Maximum CIR method. It is also shown that when the PF method is employed, the probability of user throughput of greater than 2 Mbps in the vicinity of the cell site becomes approximately 45% (5%) for L = 1-path (2-path) fading channel, while it is almost 80% (50%) when using the Maximum CIR method. Finally, we show that the average user throughput in a 2-path Rayleigh fading channel is reduced by approximately 30% compared to that in a 1-path channel due to severe multipath interference (MPI) and that the average user throughput is strongly affected by the total traffic produced within a cell, which is directly dependent on the number of users within a cell and the data size per packet call.

This paper proposes a new method to estimate the channel impulse response for Orthogonal Frequency Division Multiplexing (OFDM) mobile radio transmission. The method employs the Recursive Least Squares (RLS) algorithm so as to exploit the correlations in frequency and time domains, and can improve the estimation accuracy. It is also applicable to both the regular and scattered pilot schemes. Computer simulations demonstrate effectiveness of the method applied to the scattered pilot scheme.

We develop efficient precomputation methods of multi-scalar multiplication on ECC. We should recall that multi-scalar multiplication is required in some elliptic curve cryptosystems including the signature verification of ECDSA signature scheme. One of the known fast computation methods of multi-scalar multiplication is a simultaneous method. A simultaneous method consists of two stages; precomputation stage and evaluation stage. Precomputation stage computes points of precomputation, which are used at evaluation stage. Evaluation stage computes multi-scalar multiplication using precomputed points. In the evaluation stage of simultaneous methods, we can compute the multi-scalar multiplied point quickly because the number of additions is small. However, if we take a large window width, we have to compute an enormous number of points in precomputation stage. Hence, we have to compute an abundance of inversions, which have large computational amount. As a result, precomputation stage requires much time, as well known. Our proposed method reduces from O(22w) inversions to O(w) inversions for a window width w, using Montgomery trick. In addition, our proposed method computes uP and vQ first, then compute uP+vQ, where P,Q are elliptic points. This procedure enables us to remove unused points of precomputation. Compared with the method without Montgomery trick, our proposed method is 3.6 times faster in the case of the precomputation stage for simultaneous sliding window NAF method with window width w=3 and 160-bit scalars under the assumption that I/M=30, S/M=0.8, where I,M,S respectively denote computational amounts of inversion, multiplication and squaring on a finite field.

In an order-specified multisignature scheme, one can verify not only a set of signers who have signed the message but also its signing order. Though we have seen several schemes with such properties proposed, none of them is given the security proof against active adversaries. The scheme can be easily modified to be an order-specified multisignature scheme, but still has the restriction that the possible signing orders are only ones of the type of serial signing. In this paper, we propose the first order-specified multisignature scheme, which is shown to be secure against adaptive chosen-message insider attacks in the random oracle model, and which allows the signing orders to form like any series-parallel graphs. The security is shown by using ID-reduction technique, which reduces the security of multisignature schemes to those of multi-round identification schemes. Furthermore, we discuss the efficiency of the proposed scheme and the upper bound of the possible number of participating signers.

Software obfuscation is a promising approach to protect intellectual property rights and secret information of software in untrusted environments. Unfortunately previous software obfuscation techniques share a major drawback that they do not have a theoretical basis and thus it is unclear how effective they are. Therefore we propose new software obfuscation techniques in this paper. The techniques are based on the difficulty of interprocedural analysis of software programs. The essence of our obfuscation techniques is a new complexity problem to precisely determine the address a function pointer points to in the presence of arrays of function pointers. We show that the problem is NP-hard and the fact provides a theoretical basis for our obfuscation techniques. Furthermore, we have already implemented a prototype tool that obfuscates C programs according to our proposed techniques and in this paper we describe the implementation and discuss the experiments results.

In order to realize a future seamless high-speed road-vehicle communication system, we have proposed using code division multiplexing (CDM) radio transmission scheme by using cyclic shifted-and-extended (CSE) codes as spread codes. As the CSE codes are generated by cyclically shifting and extending a conventionally used code, the number of codes generated from a code is limited to the length of the shift interval and the tolerable period of delayed waves also depends on the length. In this paper, based on CSE codes, we propose a method to minimize the length of the shift interval and a cancellation technique with a simple calculation in order to eliminate the interference from delayed waves caused by the reduction of the length of shift interval. The concept and the BER performances in AWGN, two-paths, and multi-path fading environments are described in this paper. As a result, the maximum transmission rate of CSE-based-CDM transmission per one-code using the newly proposed transmission scheme is 3 times as large as that using conventional CSE codes and DQPSK-CDM transmission scheme.