Showing papers in "IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences in 1991"

Plane wave diffraction by a finite sinusoidal grating

Abstract: In Chapter 3, we shall consider the same grating geometry as in Chapter 2, and analyze the diffraction problem for the H-polarized plane wave incidence. The method of solution presented in Chapter 3 for H polarization is similar to, but more complicated than, the analysis carried out for the E-polarized case. Assuming that the corrugation amplitude of the grating is small compared with the wavelength, the original problem is replaced by the problem of the H-polarized plane wave diffraction by a flat strip with a certain mixed boundary condition. In Section 3.2, the approximate boundary condition thus derived contains both the firstand second-order derivatives of the unknown scattered field, and this makes the problem formulation more complicated than in the E-polarlized case. It is to be noted that, in the E polarization, the approximate boundary condition is expressed in terms of the first-order derivatives of the scattered field. Using the approximate boundary condition for the H-polarlized case, we expand the unknown scattered field into a perturbation series and separate the problem into the zero-order and the first-order boundary value problems. In Section 3.3, 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 zeroand first-order Wiener-Hopf equations, In Section 3.4, we shall apply the factorization and decomposition procedure to obtain the exact solutions of the zeroand first-order Wiener-Hopf equations. However, the solution is formal in the sense that branch-cut integrals with unknown integrands are involved. By using a rigorous asymptotic method established recently by Kobayashi [47], we have derived high-frequency asymptotic expansions of these branch-cut integrals, and subsequently obtained asymptotic solutions of the Wiener-Hopf equations for the width of the grating large compared with the wavelength. Subsequent Section 3.5 discusses the derivation of the zeroand first-order scattered field. Taking the Fourier inverse of the solution in the transform domain and applying the saddle point method, the scattered far field in the real space is derived. In Section 3.6, representative numerical examples of the far field intensity are shown for various physical parameters, and scattering characteristics of the grating are discussed in detail. The results are also compared with our analysis for the E-polarized case.

Radar image processing for locating underground linear objects

Abstract: This paper presents an image processing method for locating underground linear objects using ground-probing radar data. One of the problems when using groundprobing radars is how to know the microwave propagation velocity in the ground of the test area. The proposed method employs the Hough transform of the edge contours of pulse echoes in a radar image, which appear as hyperbolas, whose shape contains the velocity information. The method uses two types of Hough transformation. The first one transforms each edge position and orientation to a straight line on the transform plane that has the horizontal location of the object and the square of velocity as its two axes to rapidly estimate the velocity. The second one, which transforms each edge position to a parabola on the transform plane having the horizontal location of the object and the square of echo response time as its two axes, realizes aperture synthesis to give a focused echo source. A preliminary experiment confirms accurate depth measurement for multiple parallel pipes lying close toghether.

Processing Unknown Words in Continuous Speech Recognition

Abstract: Current continuous speech recognition systems essentially ignore unknown words. Systems are designed to recognize words in the lexicon. However, for using speech recognition systems in real applications of spoken-language processing, it is very important to process unknown words. This paper proposes a continuous speech recognition method which accepts any utterance that might include unknown words. In this method, words not in the lexicon are transcribed as phone sequences, while words in the lexicon are recognized correctly. The HMM-LR speech recognition system, which is an integration of Hidden Markov Models and generalized LR parsing, is used as the baseline system, and enhanced with the trigram model of syllables to take into account the stochastic characteristics of a language. Preliminary results indicate that our approach is very promising.

On the complexity of hyperelliptic discrete logarithm problem

Abstract: We give a characterization for the intractability of hyperelliptic discrete logarithm problem from a viewpoint of computational complexity theory. It is shown that the language of which complexity is equivalent to that of the hyperelliptic discrete logarithm problem is in NP ∩ co-AM, and that especially for elliptic curves, the corresponding language is in NP ∩ co-NP. It should be noted here that the language of which complexity is equivalent to that of the discrete logarithm problem defined over the multiplicative group of a finite field is also characterized as in NP ∩ co-NP.

Built-in self-test in a 24 bit floating point digital signal processor

Abstract: The authors describe a built-in self-test (BIST) method implemented in a 24-b floating-point digital signal processor (DSP) using pseudorandom patterns. By use of only one pair- of LFSRs (linear feedback shift registers) and 253 words of normal instruction, 95% of the functional blocks are self-tested. The number of the test vectors is 35 million. However, the entire BIST takes only 2.6 s for the test, owing to the fast machine cycle time of 75 ns. The overhead of the test hardware is only 2.0% of the die size. The evaluation results show that a BIST is very useful for computationally intensive VLSI processors, such as a DSP.<>

A two-layer self-diplexing antenna using a circularly polarized ring patch antenna

Abstract: Isolation between the transmitting and receiving antennas due to orthogonal circular polarization has been investigated for a two-layer self-diplexing antenna. The total isolation due to the polarization and the frequency separation was shown to be more than 35 dB by the experiments. The isolation due to the orthogonal circular polarization was more than 22 dB. The isolation can be estimated according to the axial ratio of far-field radiation patterns. The antennas can be used as mobile terminal antennas for satellite communications.<>