In this paper, we propose a new feature extraction method that combines both HMT-based denoising and weighted filter bank analysis for robust speech recognition. The proposed method is made up of two stages in cascade. The first stage is denoising process based on the wavelet domain Hidden Markov Tree model, and the second one is the filter bank analysis with weighting coefficients obtained from the residual noise in the first stage. To evaluate performance of the proposed method, recognition experiments were carried out for additive white Gaussian and pink noise with signal-to-noise ratio from 25 dB to 0 dB. Experiment results demonstrate the superiority of the proposed method to the conventional ones.

This paper addresses a robust supervisory control problem for uncertain timed discrete event systems (DESs) modeled as a set of some possible timed models. To avoid the state space explosion problem caused by tick transitions in timed models, the notion of eligible time bounds is presented. Based on the notion and activity (logical) models, this paper shows how the controllability condition of a given language specification is presented as a necessary and sufficient condition for the existence of a robust supervisor to achieve the specification for any timed model in the set.

Optimum wideband beam pattern synthesis methods are usually sensitive to antenna elements gain, phase and position errors. In this letter, these errors are taken into account in a constraint optimization process, and a generalized diagonal loading algorithm is obtained. Computer simulations indicate the robustness of this new method.

This paper proposes a new robust adaptive processing algorithm that is based on the extended least squares (ELS) method with running spectrum filtering (RSF). By utilizing the different characteristics of running spectra between speech signals and noise signals, RSF can retain speech characteristics while noise is effectively reduced. Then, by using ELS, autoregressive moving average (ARMA) parameters can be estimated accurately. In experiments on real speech contaminated by white Gaussian noise and factory noise, we found that the method we propose offered spectrum estimates that were robust against additive noise.

This paper presents an efficient scaling-simulation algorithm that simulates operations of the reconfigurable mesh (RM) of size n n using the mesh with multiple partitioned buses (MMPB) of size m m (m < n). The RM and the MMPB are the two-dimensional mesh-connected computers equipped with broadcasting buses. The broadcasting buses of the RM can be used to dynamically obtain various interconnection patterns among the processors during the execution of programs, while those of the MMPB are placed only to every row and column and are statically partitioned in advance by a fixed length. We show that the RM of size n n can be simulated in steps by the MMPB of size m m (m < n), where L is the number of broadcasting buses in each row/column of the simulating MMPB. Although the time-complexity of our algorithm is less efficient than that of the fastest RM scaling-simulation algorithm, the simulating model of our algorithm is the MMPB model where the bus-reconfiguration is not allowed.

This paper presents the definition and implementation design of a low power data bus encoding scheme dedicated to system on chip video architectures. Trends in CMOS technologies focus the attention on the energy consumption issue related to on-chip global communication; this is especially true for data dominated applications such as video processing. Taking into account scaling effects a novel coupling-aware bus power model is used to investigate the statistical properties of video data collected in the system bus of a reference hardware/software H.263/MPEG-4 video coder architecture. The results of this analysis and the low complexity requirements drive the definition of a bus encoding scheme called CDSPBI (Coupling Driven Separated Partial Bus Invert), optimized ad-hoc for video data. A VLSI implementation of the coding circuits completes the work with an area/delay/power characterization that shows the effectiveness of the proposed scheme in terms of global power saving for a small circuit area overhead.

This paper is concerned with improving noise-robustness of a multi-SNR multi-band speaker identification system by introducing automatic adjustment of subband likelihood recombination weights. The adjustment is performed on the basis of subband power calculated from the noise observed just before the speech starts in the input signal. To evaluate the noise-robustness of this system, text-independent speaker identification experiments were conducted on speech data corrupted with noises recorded in five environments: "bus," "car," "office," "lobby," and "restaurant". It was found that the present method reduces the identification error by 15.9% compared with the multi-SNR multi-band method with equal recombination weights at 0 dB SNR. The performance of the present method was compared with a clean fullband method in which a speaker model training is performed on clean speech data, and spectral subtraction is applied to the input signal in the speaker identification stage. When the clean fullband method without spectral subtraction is taken as a baseline, the multi-SNR multi-band method with automatic adjustment of recombination weights attained 56.8% error reduction on average, while the average error reduction rate of the clean fullband method with spectral subtraction was 11.4% at 0 dB SNR.

A new common-bus architecture with temporal and spatial parallel access capabilities under wire-resource constraint is proposed to transfer vast quantities of data between modules inside a VLSI chip. Since bus controllers are distributed into modules, the proposed bus architecture can directly transfer data from one module to another without any central bus control unit like a Direct Memory Access (DMA) controller, which enables to reduce communication steps for data transfer between modules. Moreover, when a start address and the number of block data in both source/destination modules are determined at the first step of a data-transfer scheme, no additional address setting for the data transfer is required in the rest of the scheme, which allows us to use all the wire resources as only the "data bus." Therefore, the bus function is dynamically programmed, which results in achieving high throughput of bus communication. For example, in case of a 64-line common bus, it is evaluated that the maximum data throughput in the proposed architecture with dynamic bus-function programming is four times higher than that in the conventional DMA bus architecture with fixed 32-bit-address/32-bit-data buses.

Some conventional beamformers require the direction of the desired signal. The performance of such beamformers can substantially be degraded even in the presence of small error on the directional information. In this letter, we propose a prefilter-type beamforming scheme robust to directional error by employing a simple compensator. The performance of the proposed scheme is verified by computer simulation.

Energy consumption is one of the most critical constraints in the design of portable embedded systems. This paper describes an empirical study about the impacts of compiler optimizations on the energy consumption of the address bus between processor and instruction memory. Experiments using a number of real-world applications are presented, and the results show that transitions on the instruction address bus can be significantly reduced (by 85% on the average) by the compiler optimizations together with bus encoding.

In this paper, we propose a robust adaptive algorithm for impulsive noise suppression. The perturbation of the input signal as well as the perturbation of the estimation error are restricted by M-estimation. The threshold used in M-estimation is obtained from the proposed adaptive variance estimation. Simulations show that the proposed algorithm is less vulnerable to the impulsive noise than the conventional algorithm.

This paper describes a noise robustness technique that normalizes the cepstral amplitude range in order to remove the influence of additive noise. Additive noise causes speech feature mismatches between testing and training environments and it degrades recognition accuracy in noisy environments. We presume an approximate model that expresses the influence by changing the amplitude range and the DC component in the log-spectra. According to this model, we propose a cepstral amplitude range normalization (CARN) that normalizes the cepstral distance between maximum and minimum values. It can estimate noise robust features without prior knowledge or adaptation. We evaluated its performance in an isolated word recognition task by using the Noisex92 database. Compared with the combinations of conventional methods, the CARN could improve recognition accuracy under various SNR conditions.

A novel public watermarking algorithm based on chaotic properties is proposed. Thanks to good randomness and easy reproducibility of chaos, firstly the watermark is permuted by chaotic sequences, then a small number of reference points are selected randomly in the middle frequency bands of DCT domain, and the variable number disorder watermark bits are embedded into the neighborhood of each reference point according to chaotic sequences. The experimental results show the invisibility and robustness.

Recently with the advent of the IT and the wide spread use of the Internet, new user oriented production and logistic systems, such as the Supply Chain Management System, have been required in order to cope with the drastic and continuous changes on the markets and users' preferences. Therefore, heterogeneous database systems need to be integrated in a common environment which can cope with the heterogeneous requirements of each company under an ever-evolving changing environment. That is assurance. Autonomous Decentralized Database System (ADDS) is proposed as a system architecture in order to realize assurance in distributed database systems. In this system architecture, a loosely-consistency management technology is proposed in order to maintain the consistency of the system, each database can update autonomously, and confer the real time property. A background coordination technology, performed by an autonomous mobile agent, is devised to adapt the system to evolving situations. The system can achieve real time by allocating the information in advance among the sites that has different time constraints for updating. Moreover, an assurance information allocation technology is proposed when considering that a failure in the background coordination mechanism may lead to loss of data and unavailability of the system. This mechanism, in which the mobile agent autonomously regulate its own capacity for allocating the information, is proposed based on the real-time property and system's availability considerations. The effectiveness of the proposed architecture and technologies are evaluated by simulation.

Grid computing can help to promote high-performance computing at a low overall cost by encouraging research centers to share their resources. However, research staff usually finds it quite hard to use Grids effectively, due to the need of installing and managing new Grid software. Thus, Grid portals are created, making it easier to take advantage of the full capability of the Grid, favoring in this way its use. The goal of this paper is to describe the process of design and implementation of a Grid Portal with the aim of both supporting distributed high-performance resources and make its use by researchers as transparent as possible. This portal uses standard Grid and Web technologies. We have designed the portal so that it can be adapted to different existing Grid infrastructures, based on the Globus Toolkit, and new functionalities can be easily added. The first prototype of the portal has been tested on an experimental Grid platform, and we present encouraging experiences carried out there.

The present paper proposes differential transmission line structures on Si ULSI. Interconnect structures are examined using numerical results from a two-dimensional electromagnetic simulation (Ansoft, 2D Extractor). The co-planar and diagonal-pair lines are found to have superior characteristics for gigahertz signal propagation through long interconnects. The proposed diagonal-pair line can reduce the crosstalk noise and interconnect resource concurrently.

We had previously proposed a fuzzy logic-based buffer management scheme (BMS) called fuzzy early detection (FED), which was designed to improve transmission control protocol (TCP) performance over the unspecified bit rate (UBR) service of asynchronous transfer mode (ATM) networks. Since a weakness in FED was discovered later, we present a refined version of it named FED+ here. Maintaining the design architecture and the algorithm of FED, FED+ further adopts a specific per virtual connection accounting algorithm to achieve its design goals. The effects of TCP implementation, TCP maximum segment size, switch buffer size and network propagation delay on FED+ performance are studied through simulation. Its performance is then compared with those of pure early packet discard (EPD), P-random early detection (P-RED) and FED. Our evaluations show that FED+ is superior to the others if the issues of efficiency, fairness, robustness, buffer requirement and the ease of tuning control parameters of a BMS are considered collectively.

This paper proposes a noise robust speech recognition method using prosodic information. In Japanese, the fundamental frequency (F0) contour represents phrase intonation and word accent information. Consequently, it conveys information about prosodic phrases and word boundaries. This paper first describes a noise robust F0 extraction method using the Hough transform, which achieves high extraction rates under various noise environments. Then it proposes a robust speech recognition method using multi-stream HMMs which model both segmental spectral and F0 contour information. Speaker-independent experiments are conducted using connected digits uttered by 11 male speakers in various kinds of noise and SNR conditions. The recognition error rate is reduced in all noise conditions, and the best absolute improvement of digit accuracy is about 4.5%. This improvement is achieved by robust digit boundary detection using the prosodic information.

In mobile multimedia environment, it is very important to minimize handoff latency due to mobility. In terms of reducing handoff latency, Hierarchical Mobile IPv6 (HMIPv6) can be an efficient approach, which uses a mobility agent called Mobility Anchor Point (MAP) in order to localize registration process. However, MAP can be a single point of failure or performance bottleneck. In order to provide mobile users with satisfactory quality of service and fault-tolerant service, it is required to cope with the failure of mobility agents. In, we proposed Robust Hierarchical Mobile IPv6 (RH-MIPv6), which is an enhanced HMIPv6 for fault-tolerant mobile services. In RH-MIPv6, an MN configures two regional CoA and registers them to two MAPs during binding update procedures. When a MAP fails, MNs serviced by the faulty MAP (i.e., primary MAP) can be served by a failure-free MAP (i.e., secondary MAP) by failure detection/recovery schemes in the case of the RH-MIPv6. In this paper, we investigate the comparative study of RH-MIPv6 and HMIPv6 under several performance factors such as MAP unavailability, MAP reliability, packet loss rate, and MAP blocking probability. To do this, we utilize a semi-Markov chain and a M/G/C/C queuing model. Numerical results indicate that RH-MIPv6 outperforms HMIPv6 for all performance factors, especially when failure rate is high.

In this paper we describe a new framework of feature compensation for robust speech recognition, which is suitable especially for small devices. We introduce Delta-cepstrum Normalization (DCN) that normalizes not only cepstral coefficients, but also their time-derivatives. Cepstral Mean Normalization (CMN) and Mean and Variance Normalization (MVN) are fast and efficient algorithms of environmental adaptation, and have been used widely. In those algorithms, normalization was applied to cepstral coefficients to reduce the irrelevant information from them, but such a normalization was not applied to time-derivative parameters because the reduction of the irrelevant information was not enough. However, Histogram Equalization (HEQ) provides better compensation and can be applied even to the delta and delta-delta cepstra. We investigate various implementation of DCN, and show that we can achieve the best performance when the normalization of the cepstra and the delta cepstra can be mutually interdependent. We evaluate the performance of DCN using speech data recorded by a PDA. DCN provides significant improvements compared to HEQ. It is shown that DCN gives 15% relative word error rate reduction from HEQ. We also examine the possibility of combining Vector Taylor Series (VTS) and DCN. Even though some combinations do not improve the performance of VTS, it is shown that the best combination gives the better performance than VTS alone. Finally, the advantage of DCN in terms of the computation speed is also discussed.