Vibration generated by the non-uniform road profile propagates though each tire and the suspension and finally generates structure born noise in the interior of the passenger vehicle. In this paper, the road-booming-noise which has strong correlation with the vibration signals measured at the suspension system was compensated. Active noise control of the road-booming-noise is rather difficult to achieve because of its non-stationary characteristics. CMFX LMS (Constraint Multiple Filtered-X Least Mean Square) algorithm, which can track non-stationary process rather well, is applied. Comprison of the proposed method and the conventional MFX LMS (Multiple Filtered-X Least Mean Square) algorithm is made through the hardware-in-the-loop simulation and the feasibility of the proposed method is demonstrated with the experiment.

In this paper, we present an adaptive formant tracking algorithm for speech using closed phase WRLS-VFF-VT method. The pitch synchronous closed phase methods is known to give more accurate estimates of the vocal tract parameters than the pitch asynchronous method. However the use of a pitch-synchronous closed phase analysis method has been limited due to difficulties associated with the task of accurately isolating the closed phase region in successive periods of speech. Therefore we have implemented the pitch synchronous closed phase WRLS-VFF-VT algorithm for speech analysis, especially for formant tracking. The proposed algorithm with the variable threshold(VT) can provide a superior performance in the boundary of phone and voiced/unvoiced sound. The proposed method is experimentally compared with the other method such as two channel CPC method by using synthetic waveform and real speech data. From the experimental results, we found that the block data processing techniques, such as the two-channel CPC, gave reasonable estimates of the formant/antiformant. However, the data windows used by these methods included the effects of the periodic excitation pulses, which affected the accuracy of the estimated formants. On the other hand the proposed WRLS-VFF-VT method, which eliminated the influence of the pulse excitation by using an input estimation as part of the algorithm, gave very accurate formant/bandwidth estimates and good spectral matching.

The forward propagated ultrasonic fields resulting from a circular plane or a concave transducer in layered fluid media as well as in homogeneous water are theoretically estimated by the angular spectrum method(ASMJ) combined with Rayleigh-Sommerfeld diffraction theory(RSDT), and measured by a precision 3-D scanning system with a needle-point hydrophone. To make the aliasing error negligible on the 2-D FFT in the theoretical estimation, the spatial discretization in the ASM are carefully considered for optimal selection of spatial sampling intervals and the size of discretization area. It is shown that the estimated fields agree reasonably with the measured ones.

A conceptual simplified model of Hartmann-Sprenger tube is suggested and investigated to decouple the regurgitant mode in the present paper. In spite of high nonlinearity, the acoustic behavior of this resonant tube system is dependent on wavelength and depth of the tube. The effect of forcing frequency and tube geometry on jet regurgitant mode are studied and discussed. With a conventional axisymmetric Euler code, sensitive acoustic problems are solved and validated by comparison with analytic theories.

Evaluating log-likelihood of Gaussian mixture density is major computational burden for the keyword spotting system using continuous HMM. In this paper, we employ the bucket box intersection (BBI) algorithm to reduce the computational complexity of keyword spotting. We make some modification in implementing BBI algorithm in order to increase the discrimination ability among the keyword models. According to our keyword spotting experiments, the modified BBI algorithm reduces 50% of log-likelihood computations without performance degradation, while the original BBI algorithm under the same condition reduces only 30% of log-likelihood computations.