A new method is proposed for automatically evaluating the English pronunciation quality of non-native speakers. It is assumed that pronunciation can be rated using three criteria: the static characteristics of phonetic spectra, the dynamic structure of spectrum sequences, and the prosodic characteristics of utterances. The evaluation uses speech recognition techniques to compare the English words pronounced by a non-native speaker with those pronounced by a native speaker. Three evaluation measures are proposed to rate pronunciation quality. (1) The standard deviation of the mapping vectors, which map the codebook vectors of the non-native speaker onto the vector space of the native speaker, is used to evaluate the static phonetic spectra characteristics. (2) The spectral distance between words pronounced by the non-native speaker and those pronounced by the native speaker obtained by the DTW method is used to evaluate the dynamic characteristics of spectral sequences. (3) The differences in fundamental frequency and speech power between the pronunciation of the native and non-native speaker are used as the criteria for evaluating prosodic characteristics. Evaluation experiments are carried out using 441 words spoken by 10 Japanese speakers and 10 native speakers. One half of the 441 words was used to evaluate static phonetic spectra characteristics, and the other half was used to evaluate the dynamic characteristics of spectral sequences, as well as the prosodic characteristics. Based on the experimental results, the correlation between the evaluation scores and the scores determined by human judgement is found to be 0.90.

This paper proposes and investigates an adaptive equalizer with diversity-combining over a multipath fading channel. It consists of two space-diversity antennas and a Ts/2-spaced decision-feedback-equalizer (DFE). Received signals from the two antennas are alternatively switched and fed into the feed forward-filter of DFE. We call this structure a Switched Input Combining Equalizer with diversity-combining (SICE). By using an SICE, the receiver structure for combining diversity equalization can be simplified, because it needs only two receiver sections up to IF BPF. The bit error rate (BER) performance of SICE was evaluated by both computer simulation and experiment over a multipath fading channel. We experimentally confirmed the excellent BER performance, around 1% of BER over a multipath fading channel at 160Hz of maximum doppler fading frequency. Therefore, the proposed SICE is applicable to highly reliable transmission in the 1.5-GHz-band mobile radio.

A Decision Feedback Equalizer (DFE) structure with a varying number of tap lengths was used with a recursive least squares (RLS) algorithm to determine tradeoffs between equalizer size and performance in mobile and portable digital radio systems. A mobile channel simulator, SMRCIM, was used to demonstrate how much an equalizer can improve the BER in real world urban channels. The results show that at 850MHz, the DFE is unable to improve the BER when the mobile terminal exceeds speeds of 115km/h for U.S. Digital Cellular systems. The performance of adaptive equalization for indoor high data rate systems was evaluated using the indoor channel simulator SIRCIM, and we found that DFEs have excellent performance for indoor radio channels. For simple structures, the BER is less than 10-3 at 15dB Eb/NO using coherent QPSK modulation. Finally, an equalizer structure for non-coherent π/4 DQPSK modulation was developed and simulation results are presented.

This paper proposes a new speaker adaptation method using a speaker weighting technique for multiple reference speaker training of a hidden Markov model (HMM). The proposed method considers the similarities between an input speaker and multiple reference speakers, and use the similarities to control the influence of the reference speakers upon HMM. The evaluation experiments were carried out through the/b, d, g, m, n, N/phoneme recognition task using 8 speakers. Average recognition rates were 68.0%, 66.4%, and 65.6% respectively for three test sets which have different speech styles. These were 4.8%, 8.8%, and 10.5% higher than the rates of the spectrum mapping method, and also 1.6%, 6.7%, and 8.2% higher than the rates of the multiple reference speaker training, the supplemented HMM. The evaluation experiments clarified the effectiveness of the proposed method.

This paper describes a new supervised speaker adaptation method based on vector field smoothing, for small size adaptation data. This method assumes that the correspondence of feature vectors between speakers can be viewed as a kind of smooth vector field, and interpolation and smoothing of the correspondence are introduced into the adaptation process for higher adaptation performance with small size data. The proposed adaptation method was applied to discrete HMM based speech recognition and evaluated in Japanese phoneme and phrase recognition experiments. Using 10 words as the adaptation data, the proposed method produced almost the same results as the conventional codebook mapping method with 25 words. These experiments clearly comfirmed the effectiveness of the proposed method.

Three custom LSIs for EB60, a direct wafer exposure electron beam system, have been developed using 0.8 µm BiCMOS and SST bipolar technologies. The three LSIs are i) a shot cycle control LSI for controlling each exposure cycle time, ii) a linear matrix computation LSI for coordinate modification of the exposure pattern data, and iii) a position calculation LSI for determining the precise position of the wafer. These LSIs allow the deflection corrector block of the revised EB60 to be realized on a single board. A new adaptive pipeline control technique which optimizes each shot period according to the exposure data is implemented in the shot-cycle control LSI. The position calculation LSI implements a new, highly effective 2-level pipeline exposure technique, the levels refer to major-field-deflection and minor-field-deflection. The linear-matrix computation LSI is designed not only for the EB60 but also for a wide variety of parallel digital processing applications.

We have demonstrated the successful fabrication of the monolithic integration of a GaAs metalsemiconductor field-effect transistor (MESFET), an AlGaAs/InGaAs laser and a p-n photodetector grown on a SiO2 backcoated p-Si substrate using selective regrowth by metalorganic chemical vapor deposition (MOCVD). The use of SiO2 backcoated Si substrate is effective in suppressing unintentional Si autodoping and obtaining a good pinch-off GaAs MESFET. The MESFET with 2.5400 µm2 gate exhibited a transconductance of 90 mS/mm and a threshold voltage of 2.2 V. The reliability of the laser on the Si substrate can be improved by the strain-relieved AlGaAs/InGaAs laser with the InGaAs intermediate layer. The longest lifetime of the laser is 8 h at 27. During the GaAs layer growth, the p-n photodetector is formed near the surface of the p-Si substrate by diffusing the As atoms.

The diffraction of a plane electromagnetic wave by a parallel-plate waveguide cavity with a thick planar termination is rigorously analyzed for both the E and the H polarization using the Wiener-Hopf technique. Introducing the Fourier transform for the unknown scattered field and applying boundary conditions in the transform domain, the problem is formulated in terms of the simultaneous Wiener-Hopf equations, which are solved exactly in a formal sense via the factorization and decomposition procedure. Since the formal solution involves an infinite number of unknowns and branch-cut integrals with unknown integrands, approximation procedures based on rigorous asymptotics are further presented to yield the approximate solution convenient for numerical computations. The scattered field inside and outside the cavity is evaluated by taking the inverse Fourier transform and applying the saddle point method. Representative numerical examples of the monostatic and bistatic radar cross sections are presented for various physical parameters, and the scattering characteristics of the cavity are discussed in detail.

Progress of speech recognition based on the hidden Markov model has made it possible to realize man-machine dialogue systems capable of operating in real time. In spite of considerable effort, however, few systems have been successfully developed because of the lack of appropriate dialogue models. This paper reports on some of technology necessary to develop a dialogue system with which one can converse comfortably. The emphasis is placed on the following three points: how a human converses with a machine; how errors of speech recognition can be recovered through conversation; and what it means for a machine to be cooperative. We examine the first problem by investigating dialogues between human speakers, and dialogues between a human speaker and a simulated machine. As a consideration in the design of dialogue control, we discuss the relation between efficiency and cooperativeness of dialogue, the method for confirming what the machine has recognized, and dynamic adaptation of the machine. Thirdly, we review the research on the friendliness of a natural language interface, mainly concerning the exchange of initiative, corrective and suggestive answers, and indirect questions. Lastly, we describe briefly the current state of the art in speech recognition and synthesis, and suggest what should be done for acceptance of spontaneous speech and production of a voice suitable to the output of a dialogue system.

A system is constructed for the processing of question-answer dialogue as a subsystem of the speech response device. In order to increase the adaptability to dialogue topics, rules for dialogue processing are classified into three groups; universal rules, topic-dependent rules and task-dependent rules, and example-based description is adopted for the second group. The system is disigned to operate only with information on the content words of the user input. As for speech synthesis, a function is included in the system to control the focal position. Introduction and guidance of ski areas are adopted as the dialogue domain, and a prototype system is realized on a computer. The dialogue example performed with the prototype indicates the propriety of our method for dialogue processing.

An adaptive fuzzy network (AFN) is described that can be used to implement most of fuzzy logic functions. We introduce a learning algorithm largely borrowed from backpropagation algorithm and train the AFN system for several typical fuzzy problems. Simulations show that an adaptive fuzzy network can be implemented with the proposed network and algorithm, which would be impractical for a conventional fuzzy system.

This paper describes the application of adaptive filter and wave equalization technology to acoustics, and to noise reduction of a machine using acoustic field control. Firstly, some problems inherent in applying active noise control (ANC) technology to noise reduction in consumer products are pointed out. In particular, the behavior of Error-Adaptive Control, as named by the authors, is analyzed precisely. Secondly, the relationship between coherence and the performance of active control is investigated. The fact that coherence is large or small is more effective for ANC when adaptive control is used rather than fixed-coefficient filter control. The effects of sound spatial coherence on adaptive ANC are studied precisely. The study looks into the relationship between minimum mean square error and input signal variance, or coherence, which has been measured previously. In three-dimensional spatial control, several microphones and speakers are needed for ANC, and several acoustic paths are present. ANC performance in three-dimensional space was evaluated by multiple coherence, which shows the degree of multiple spatial correlation. Thirdly, the paper describes the application of ANC technology to compressor noise in a refrigerator, a mass product. The problem was solved by treating the machine chamber as a one-dimensional duct, preventing howl, and using Error-Adaptive control. The second application is to fan noise in a small device. The authors discovered that the spatial coherence of the sound is low in the vicinity of the fan. This causes ANC to operate at a low level.

This paper proposes a new adaptive algorithm for acoustic echo cancellers with four times the convergence speed for a speech input, at almost the same computational load, of the normalized LMS (NLMS). This algorithm reflects both the statistics of the variation of a room impulse response and the whitening of the received input signal. This algorithm, called the ESP (exponentially weighted step-size projection) algorithm, uses a different step size for each coefficient of an adaptive transversal filter. These step sizes are time-invariant and weighted proportional to the expected variation of a room impulse response. As a result, the algorithm adjusts coefficients with large errors in large steps, and coefficients with small errors in small steps. The algorithm is based on the fact that the expected variation of a room impulse response becomes progressively smaller along the series by the same exponential ratio as the impulse response energy decay. This algorithm also reflects the whitening of the received input signal, i.e., it removes the correlation between consecutive received input vectors. This process is effective for speech, which has a highly non-white spectrum. A geometric interpretation of the proposed algorithm is derived and the convergence condition is proved. A fast profection algorithm is introduced to reduce the computational complexity and modified for a practical multiple DSP structure so that it requires almost the same computational load, 2L multiply-add operations, as the conventional NLMS. The algorithm is implemented in an acoustic echo canceller constructed with multiple DSP chips, and its fast convergence is demonstrated.

This paper proposes a new adaptive algorithm of the FIR type digital filter for an acoustic echo canceller and similar application fields. Unlike an echo canceller for line, an acoustic echo canceller requires a large number of taps, and it must work appropriately while it is driven by colored input signal. By controlling the filter tap length and updating filter coefficients multiple times during a single sampling interval, the proposed algorithm improves the convergence characteristics of adaptation even if colored input signal is introduced. This algorithm is maned VT-LMS after variable tap length LMS. The results of simulation show the effectiveness of the proposed algorithm not only for white noise but also for colored input signal such as speech. The VT-LMS algorithm has better convergence characteristice with very little extra computational load compared to the conventional algorithm.

This paper relates to a novel algorithm for fast estimation of the coefficients of the adaptive FIR filter. The novel algorithm is derived from a first order IIR filter experssion clarifying the estimation process of the NLMS (normalized least mean square) algorithm. The expression shows that the estimation process is equivalent to a procedure extracting the cross-correlation coefficient between the input and the output of an unknown system to be estimated. The interpretation allows to move a subtraction of the echo replica beyond the IIR filter, and the movement gives a construction with the IIR filter coefficient of unity which forms the arithmetic mean. The construction in comparison with the conventional NLMS algorithm, improves the covergence rate extreamly. Moreover, when we use the construction with a simple technique which limits the term of calculating the correlation coefficient in the beginning of a convergence process, the convergence delay becomes negligible. This is a very desirable performance for acoustic echo canceller. In this paper, double-talk and echo path fluctuation are also studied as the first stage for application to acoustic echo canceller. The two subjects can be resolved by introducing two switches and delays into the evaluation process of the correlation coefficient.

An acoustic echo canceller that also cancels room noise is proposed. This system has an additive (noise reference) input port, and a noise canceller (NC) precedes the echo canceller (EC) in a cascade configuration. The adaptation control problem for the cascaded echo and noise canceller is solved by controlling the adaptation process to match the occurrence of intermittent speech/echo; the room noise is a stationary signal. A simulation shows that adaptation using the NLMS algorithm is very effective for the echo and noise cancellation. Sub-band cancelling techniques are utilized. Noise cancellation is realized with a lower band EC. Hardware is implemented and its performance evaluated through experiments under a real acoustic field. The combination of the EC with NC maintains excellent performance at all echo to room noise power ratios. It is shown that the proposed canceller overcomes the disadvantages traditionally associated with ECs and NSc.

The detection problem of fluctuating radar targets in the presence of interference (noise and clutter) is considered; the assumed model for both target and clutter is a zero-mean stationary Gaussian random process with assigned power spectral densities. The pertaing optimum linear processor, namely the Optimized Filtering, is derived and its performance are evaluated in different operating conditions, including mismatching with the designed model. Finally, comparison with filtering techniques designed for targets with zero spectral width, i.e. the Moving Target Detector, are performed.

This paper presents a method for LSI implementation of a long-tap acoustic echo canceller algorithm using the residue number system (RNS) and the mixed-radix number system (MRS). It also presents a quantitative comparison of echo canceller architectures, one using the RNS and the other using the binary number system (BNS). In the RNS, addition, subtraction, and multiplication are executed quickly but scaling, overflow detection, and division are difficult. For this reason, no echo canceller using the RNS has been implemented. We therefore try to design an echo canceller architecture using the RNS and the NLMS algorithm. It is shown that the echo canceller algorithm can be effectively implemented using the RNS by introducing the MRS. The quantitative comparison of echo canceller architectures shows that a long-tap acoustic echo canceller can be implemented more effectively in terms of chip size and power dissipation by the architecture using the RNS.

Most communication cables are laid underground. In order to make construction and maintenance works easier, systems to detect buried objects have already been developed using the electromagnetic pulse radar technique. However, existing detection systems are not really practical due to their rather limited processing speed. To achieve sufficient processing speed, two dedicated custom FFT LSI's are designed and realized with 0.8 µm-CMOS technology. The two chips have an equivalent processing capacity of 200 MOPS. An efficietn hardware algorithm for address generation and 2 word parallel processing are introduced. In addition, an enhanced system organization is developed together with an improved pattern recognition scheme and aperture synthesis hardware. The new processor executes a FFT/parameter extraction operation in 4 seconds and aperture synthesis in 1 second. This speed meets the design target, and a real time detection system for underground objects becomes possible.

Recently most of the singularities of the equivalent edge currents for flat plates were eliminated by the authors using the paths of most rapid phase variation. A unique direction on the plate was determined for given incidence and observer. This paper extends this method for arbitrary angle wedges and presents the new expressions of the equivalent edge currents. The resultant expressions are valid for any incidence and observation aspects and have no false singularities. Diffraction patterns and radar cross sections of 3-D objects composed of wedges are calculated by using these currents. They show good agreements with experimental data or the results by the other methods.