This paper addresses the system throughput maximization problem for HAPS third generation cellular systems. We assume that the Stratospheric Platform is able to perform a perfect link gain estimation for all mobile terminals, such that a centralized resource allocation strategy is made possible. A classical 3G wireless scenario is considered, where traffics characterized by different bit rates coexist with Best Effort Traffic services without stringent bit rate constraints. In this scenario, we firstly envisage three Rate Assignment schemes for best effort terminals which aim at achieving the maximum system throughput subject to different bit rate constraints. For the second envisaged rate assignment scheme, which represents the best compromise between service fairness and throughput, we then propose a simplified approach that allows to noticeably decrease the implementation complexity with a slight performance degradation.

In this paper, a novel confidence scoring method that is applied to N-best hypotheses (word candidates) output from an HMM-based classifier is proposed. In the first pass of the proposed method, the HMM-based classifier with monophone models outputs N-best hypotheses and boundaries of all monophones in the hypotheses. In the second pass, an SM (Subspace Method)-based verifier tests the hypotheses by comparing confidence scores. To test the hypotheses, at first, the SM-based verifier calculates the similarity between phone vectors and an eigen vector set of monophones, then this similarity score is converted into a likelihood score with normalization of acoustic quality, and finally, an HMM-based likelihood of word level and an SM-based likelihood of monophone level are combined to formulate the confidence measure. Two kinds of experiments were performed to evaluate this confidence measure on speaker-independent word recognition. The results showed that the proposed confidence scoring method significantly reduced the word error rate from 4.7% obtained by the standard HMM classifier to 2.0%, and in an unknown word rejection, it reduced the equal error rate from 9.0% to 6.5%.

The present paper deals with the blind deconvolution of a Multiple-Input Multiple-Output Finite Impulse Response (MIMO-FIR) system. To deal with the blind deconvolution problem using the second-order statistics (SOS) of the outputs, Hua and Tugnait considered it under the conditions that a) the FIR system is irreducible and b) the input signals are spatially uncorrelated and have distinct power spectra. In the present paper, the problem is considered under a weaker condition than the condition a). Namely, we assume that c) the FIR system is equalizable by means of the SOS of the outputs. Under b) and c), we show that the system can be blindly identified up to a permutation, a scaling, and a delay using the SOS of the outputs. Moreover, based on this identifiability, we show a novel necessary and sufficiently condition for solving the blind deconvolution problem, and then, based on the condition, we propose a new algorithm for finding an equalizer using the SOS of the outputs, while Hua and Tugnait have not proposed any algorithm for solving the blind deconvolution under the conditions a) and b).

In this paper we consider blind source separation (BSS) problem of signals which are spatially uncorrelated of order four, but temporally correlated of order four (for instance speech or biomedical signals). For such type of signals we propose a new sufficient condition for separation using fourth order statistics, stating that the separation is possible, if the source signals have distinct normalized cumulant functions (depending on time delay). Using this condition we show that the BSS problem can be converted to a symmetric eigenvalue problem of a generalized cumulant matrix Z(4)(b) depending on L-dimensional parameter b, if this matrix has distinct eigenvalues. We prove that the set of parameters b which produce Z(4)(b) with distinct eigenvalues form an open subset of RL, whose complement has a measure zero. We propose a new separating algorithm which uses Jacobi's method for joint diagonalization of cumulant matrices depending on time delay. We empasize the following two features of this algorithm: 1) The optimal number of matrices for joint diago- nalization is 100-150 (established experimentally), which for large dimensional problems is much smaller than those of JADE; 2) It works well even if the signals from the above class are, additionally, white (of order two) with zero kurtosis (as shown by an example).

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.

A fast nearest neighbor codeword search algorithm for vector quantization (VQ) is introduced. The algorithm uses three significant features of a vector, that is, the mean, the variance and the norm, to reduce the search space. It saves a great deal of computational time while introducing no more memory units than the equal-average equal-variance codeword search algorithm. With two extra elimination criteria based on the mean and the variance, the proposed algorithm is also more efficient than so-called norm-ordered search algorithm. Experimental results confirm the effectiveness of the proposed algorithm.

As a typical planar antenna in Japan, a microstrip antenna and radial line slot antenna are chosen and some original technologies are introduced for them. About the microstrip antenna, the analyzing method is described first and the method based on the theory of microstrip planar circuit born in Japan is introduced. According to the formulas derived by this method, the design procedure considering the bandwidth is established. In addition, it is shown clearly that a microstrip antenna can produce the circular polarizations at two kinds of frequencies with a single feed. Furthermore, two kinds of broadband techniques born in Japan are picked up. About other unique microstrip antennas, they may be introduced in a suitable section each time. As for the RLSA, the history on invention is briefly presented. The radiation mechanisms depending on the slot-set arrangement and the excitation mode are discussed. The slot-coupling analysis to simulate the excitation of a two-dimensional uniformly-excited slot array is explained. The simple design based on the operation with traveling-wave propagation is also described. The technical progress to keep high efficiency in a wide gain range for satellite-TV reception is reviewed. Extensions of the RLSAs to millimeter-wave bands and plasma etching systems are finally summarized.

The problem of separating blindly independent sources from a convolutive mixture cannot be addressed in its widest generality without resorting to statistics of order higher than two. The core of the problem is in fact to identify the paraunitary part of the mixture, which is addressed in this paper. With this goal, a family of statistical contrast is first defined. Then it is shown that the problem reduces to a Partial Approximate Joint Diagonalization (PAJOD) of several cumulant matrices. Then, a numerical algorithm is devised, which works block-wise, and sweeps all the output pairs. Computer simulations show the good behavior of the algorithm in terms of Symbol Error Rates, even on very short data blocks.

This paper presents a hardware-efficient architecture of tree-depth scan (TDS) and multiple quantization (MQ) scheme for zerotree coding in MPEG-4 still texture coder. The proposed TDS architecture can achieve its maximal throughput to area ratio and minimize the external memory access with only one wavelet-tree size on-chip buffer. The MQ scheme adopts the power-of-two (POT) quantization to realize a cost-effective hardware implementation. The prototyping chip has been implemented in TSMC 0.35 µm CMOS 1P4M technology. This architecture can handle 30 4-CIF (704576) frames per second with five spatial scalability and five SNR scalability layers at 100 MHz working frequency.

Digital halftoning is a technique to convert a continuous-tone image into a binary image consisting of black and white dots. It is an important technique for printing machines and printers to output an image with few intensity levels or colors which looks similar to an input image. This paper surveys how algorithm engineering can contribute to digital halftoning or what combinatorial problems are related to digital halftoning. A common criterion on optimal digital halftoning leads to a negative result that obtaining an optimal halftoned image is NP-complete. So, there are two choices: approximation algorithm and polynomial-time algorithm with relaxed condition. Main algorithmic notions related are geometric discrepancy, matrix (or array) rounding problems, and network-flow algorithms.

A term rewriting system which effectively preserves recognizability (EPR-TRS) has good mathematical properties. In this paper, a new subclass of TRSs, layered transducing TRSs (LT-TRSs) is defined and its recognizability preserving property is discussed. The class of LT-TRSs contains some EPR-TRSs, e.g., {f(x)f(g(x))} which do not belong to any of the known decidable subclasses of EPR-TRSs. Bottom-up linear tree transducer, which is a well-known computation model in the tree language theory, is a special case of LT-TRS. We present a sufficient condition for an LT-TRS to be an EPR-TRS. Also reachability and joinability are shown to be decidable for LT-TRSs.

In this paper, parallelization methods for quantum circuits are studied, where parallelization of quantum circuits means to reconstruct a given quantum circuit to one which realizes the same quantum computation with a smaller depth, and it is based on using additional bits, called ancillae, each of which is initialized to be in a certain state. We propose parallelization methods in terms of the number of available ancillae, for three types of quantum circuits. The proposed parallelization methods are more general than previous one in the sense that the methods are applicable when the number of available ancillae is fixed arbitrarily. As consequences, for the three types of n-bit quantum circuits, we show new upper bounds of the number of ancillae for parallelizing to logarithmic depth, which are 1/log n of previous upper bounds.

An automated synthesis for analog computational circuits in transistor-level configuration is presented. A cell-based structure is introduced to place moderate constraints on the MOSFET circuit topology. Even though each cell has a simple structure that consists of one current path with four transistors, common analog building blocks can be implemented using combinations of the cells. A genetic algorithm is applied to search circuit topologies and transistor sizes that satisfy given specifications. Synthesis capabilities are demonstrated through examples of three types of computational circuits; absolute value, squaring, and cubing functions by using computer simulations and real hardware.

In the FDTD simulation of microwave circuits, a device in very small size compared with the wavelength is often handled as a lumped element, but it may still occupy more than one cell instead of a wire structure without volume routinely employed in classical extended FDTD algorithms. In this paper, two modified extended FDTD algorithms incorporating a lumped element occupying more than one cell are developed directly from the integral form of Maxwell's equations based on the assumption whether displacement current exists inside the region where a device is present. If the displacement current exists, the modified extended FDTD algorithm can be represented as a Norton equivalent current-source circuit, or otherwise as a Thevenin equivalent voltage-source circuit. These algorithms are applied in the microwave line loaded by a lumped resistor and an active antenna to illustrated the efficiency and difference of the two algorithms.

A dual-polarized multimode CDMA based local multipoint distribution system (LMDS) is presented. The twisted sector concept and narrowed sector cell are proposed to improve the system performance. Inter-cell interference is analyzed and discussed for the downstream direction based on hexagonal cell architecture. The bit error rate (BER) performance of a multimode CDMA scheme is investigated in terms of the worst case for high order modulation. The simulation results show that the proposed cell structure obtains better power efficiency and makes the multimode CDMA scheme feasible in LMDS deployment.

Existing research related to RSVP with mobility support has mainly focused on maintaining reservation state along the routing path, which changes continuously with the movements of mobile host (MH), without much overhead and delay. However, problems such as deepening RSVP's inherent scalability problem and requiring significant changes in the existing network infrastructure have not been adequately addressed. In this paper, we propose a new approach, known as Concatenation and Optimization for Reservation Path (CORP), which addresses these issues. In CORP, each BS pre-establishes pseudo reservations to its neighboring BSs in anticipation of the MH's movement. When the MH moves into another wireless cell, the associated pseudo reservation is activated and concatenated to the existing RSVP session to guarantee continuous QoS support. Because a pseudo reservation is recognized as a normal RSVP session by intermediate routers, little change is required in the current Internet environment to support both movements within a single routing domain and between two different routing domains. CORP also dynamically optimizes the extended reservation path to avoid the infinite path extension problem. Multicast addressing is used to further reduce resource consumption in the optimization process. The experimental results of the CORP implementation demonstrate that it significantly reduces the delay and overhead caused by handoffs compared to the case of establishing a new RSVP session. The improvement increases as the distance between the MH and its correspondent host (CH) grows.

We propose a new phase interpolator that provides precise fractional phase pulses without the need to adjust circuit constants. The variable phases are produced by detecting the coincidence of two voltages, the ramp wave and the threshold voltage. The new phase interpolator can keep the same ramp wave slope and the same threshold voltage for different output phases. This significantly reduces the power dissipation of the voltage comparator. This phase interpolator can be applied to various timing circuits and clock generators, such as frequency multipliers and direct digital synthesizers. We present a novel frequency doubler, a novel frequency tripler, a direct digital synthesizer (DDS), and a novel wideband DDS (WDDS) as applications of our new phase interpolator, which uses 0.35-mm CMOS process technology. Experimental results confirm the functionarity of the new phase interpolator. An 8-bit complete DDS IC dissipates only 2.1 mA at a 50-MHz clock rate and a supply voltage of 2.8 V.

A new analog equalizer supporting 10Base-T central office and remote terminal (Category-3) LAN applications is developed. It provides robust 10 Mbps data transmission (10Base-T) at loop length up to 400 meters. The equalizer with high frequency gain boost capability is controlled automatically by the proposed AZC (Adaptive Zero Control) loop according to cable length. It is implemented using AMS 0.8 µm CMOS technology.

Methods usually employed for carrying out division in logic are based on algebraic or Boolean techniques. Algebraic division is fast but results may be less than optimal. Boolean division will yield better results but generally it is much slower because a minimization step is required. In [4], Kim and Dietmeyer proposed a new type of division, called extended algebraic division, and described a heuristic algorithm for it. A feature is that, unlike Boolean division, it does not require a minimization step. The present paper is concerned with an efficient algorithm for exact extended algebraic division. The algorithm was developed within the SIS environment, a program for logic synthesis developed at U.C. Berkeley. Experiments on factoring PLA's demonstrate a significant improvement in quality with a reasonable increase in run time.

Approximately synchronized CDMA (AS-CDMA) can reduce the inter-channel interference in a cell to zero. This property of AS-CDMA is an advantage over the conventional DS-CDMA. However, the inter-cell interference of the AS-CDMA cellular system has not been sufficiently examined previously. Therefore, the synthetic performance of AS-CDMA cellular system also has not been sufficiently clarified previously. Some factors that affect the inter-cell interference of the AS-CDMA cellular system were theoretically examined, and evaluated by using computer simulation. As the result, we found that transmission power control is effective for reducing the inter-cell interference of the AS-CDMA cellular system. In addition, the synthetic performance of AS-CDMA cellular system was clarified for the first time. Consequently, it was also found that the synthetic performance of the AS-CDMA cellular system is higher than that of the conventional DS-CDMA cellular system.