Zero-crossings with Peak amplitudes (ZCPA) model motivated by human auditory periphery was proposed to extract reliable features form speech signals even in noisy environments for robust speech recognition. In this paper, the performance of the ZCPA model is further improved by incorporating conventional speech processing techniques into the model output. Spectral and cepstral representations of the ZCPA model output are compared, and the incorporation of dynamic features with several different lengths of time-derivative window are evaluated. Also, comparative evaluations with other front-ends in real-world noisy environments are performed, and result in the superiority of the ZCPA model.

In this paper, we describe the one channel five-way, V/U/M/N/S (Voice/Unvoice/Nasal/Silent), classification algorithm for automatically classifying speech. The decision making process is viewed as a pattern viewed as a pattern recognition problem. Two aspects of the algorithm are developed: feature selection and classifier type. The feature selection procedure is studied for identifying a set of features to make V/U/M/N/S classification. The classifiers used are a vector quantization (VQ), a neural network(NN), and a decision tree method. Actual five sentences spoken by six speakers, three male and three female, are tested with proposed classifiers. From a set of measurement tests, the proposed classifiers show fairly good accuracy for V/U/M/N/S decision.

This paper proposes a simple method to measure system's performance in target tracking problems. Essentially employing the Cramer-Rao Lower Bound (CRLB) on tracking accuracy, an algorithm of predicting system's performance under various scenarios is developed. The input data is a collection of measurements over tim from sensors embedded in Gaussian noise. The target of interest may not maneuver over the processing time interval while the own ship observing platform may maneuver in an arbitrary fashion. The proposed approach is demonstrated and discussed through simulation results.

The validity of the image model is investigated both analytically and experimentally in a half space with an infinite single reflecting surface present. This paper exploits the Sommerfeld integral that represents the exact solution for the reflected field in the half space. The solution is shown to be obtained by direct numerical integration which yields more accurate and stable results. The predicted results from the image model are compared to those from the direct numerical integration of the Sommerfeld integral. It is also experimentally demonstrated that the image model gives acceptably accurate results. It is of significance that this paper reveals analytical and experimental validation of using the image model except near-grazing incidence.

Transient acoustic and elastic wave propagation in inhomogeneous media are studied by using the Fourier method. It is known that the fourier method has advantages in memory requirements and computing speed over conventional methods such as FDM and FEM, because the Fourier method needs less grid points for achieving the same accuracy. To verify the proposed numerical scheme, several examples having analytic solutions are considered, where two different semi-infinite media are in contact along a plane boundary. The comparisons of numerical results by the Fourier method and analytic solutions show good agreements. In addition, the fourier method is applied to a layered half-plane, in which an elastic semi-infinite medium is covered by an elastic layer of finite thickness. It is showed how to derive the analytic solutions by using the Cagniard-de Hoop method. The numerical solutions are in excellent agreements with analytic results.

Many of the active noise control system utilize a form of the least mean square(LMS) algorithm. This paper discusses the dependence of the convergence rate on the acoustic error path in the popular algorithm which is conventional "filtered-x LMS" and introduces new algorithm "constrained filtered-x LMS". The proposed method increase the convergence region regardless of the time-delay in the acoustic error path. In the algorithms, coefficients of the controller are adapted using the residuals of constrained structure which are defined in such a way that the control process become stationary. Advantages of constrained filtered-x LMS algorithm is illustrated by convergence analysis in the mean sense.

This paper prolposes a VFM (Variable Fading Memory)Kalman filtering and applies it to the parameter estimation for time-varying signals. By adaptively calculating the fading memory, the proposed algorithm does not require any predetermined fading memory when estimating the time-varying signal parameter. Moreover, the proposed algorithm has faster convergence speed than fixed fading memory one in case the signal contains an impulsive outlier. The performance of parameter estimation for time-varying signal is evaluated by computer simulation for two cases, one of which is the chirp signal whose frequency varies linearly with time and the other is the chip signal with an impulsive outlier. The experimental results show that the VFM Kalman filtering estimates the parameter of the chirp signal more rapidly than the fixed fading memory one in the region of an outlier.

Spectral information at a speech coder should be quantized with sufficient accuracy to keep perceptually transparent output speech. Spectral information at a low bit rate speech coder is usually transformed into corresponding line spectrum pair parameters and is often quantized with a vector quantization algorithm. As the vector quantization algorithm generally has high complexity in the optimal code vector searching routine, the complexity reduction in that routine is investigated using the ordering property of the line spectrum pair. When the proposed complexity reduction algorithm is applied to the well-known split vector quantization algorithm, the 46% complexity reduction is achieved in the distortion measure compu-tation.

In this paper, we analyze the effects of interference canceller on the performance of multitone DS/CDMA system proposed by Vandendorpe[5]. There are various kinds of interference canceller suggested by different researchers including parallel and successive cancellers and we adopt a canceller used by Yoon et al.[9] which is a kind of parallel canceller. We consider three kinds of interferences, that is, multipath interference(MPI), interchannel interference(ICI) and multiple access interference(MAI). The ICI is the interference between multitones. The equations for variances. are derived for the inteferences and thermal noise used for signal to noise ratio calculation. We also consider RAKE reception over multipath channel which is modeled as lowpass equivalent linear filter and three stage interference canceller used for performance improvement. We show the performance results for number of canceller stage, diversity order and number of users and draw some conclusions that interference canceller is effective in multitone DS/CDMA system and the performance is further improved with the higher order of diversity and larger number of PN chips.