|
1. |
Improvement of algorithm using interval analysis for solution of nonlinear circuit equations |
|
Electronics and Communications in Japan (Part I: Communications),
Volume 70,
Issue 9,
1987,
Page 1-10
Kohshi Okumura,
Akira Kishimas,
Shuichi Saeki,
Preview
|
PDF (622KB)
|
|
摘要:
AbstractThis paper describes an improvement of the solution by Krawczyk, Moore and Jones (KMJ algorithm) for nonlinear equations based on the interval analysis. When the KMJ algorithm is applied to practical problems such as the determination of the operating point of a multistable electronic circuit, a large computation time is required. The reason for the large computation time is as follows: (i) The existence of the solution is determined using the direct‐replacement type interval extensions of functions, which enlarges the search region for the solutions; (ii) the region partitioning is repeated until a regionXis obtained such thatK(X) ⊆X, whereK(X) is Krawczyk's interval function andXis the interval region. The situation is demonstrated by an example. As a solution for these problems, the following method is proposed.(1) The given initial region is partitioned into subregions, and some subregions are excluded from consideration if the solution is judged as nonexistent by the interval extensions of functions; (2) whenK(X) ⊆X, does not apply, the regionXis partitioned intoXCδX‐ (K(X) ∩X) andXK∩K(X) ∩X, and the KMJ algorithm is applied toXCandXKto examine the existence of the solutions. Finally, as a computation example, the KMJ algorithm with the above two elaborations is applied to the nonlinear circuit equations with two to four variables. It is demonstrated that the computation time can be reduced considerably with the original
ISSN:8756-6621
DOI:10.1002/ecja.4410700901
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1987
数据来源: WILEY
|
2. |
Role of formant frequencies and bandwidths in speaker perception |
|
Electronics and Communications in Japan (Part I: Communications),
Volume 70,
Issue 9,
1987,
Page 11-21
Hisao Kuwabara,
Kengo Ohgushis,
Preview
|
PDF (762KB)
|
|
摘要:
AbstractThis paper considers the relation between the vocal tract characteristics under the personality of speech. A pitch synchronous analysis‐synthesis system was constructed in which the formant frequencies and the bandwidths can be modified independently. The speech was synthesized by controlling these parameters in natural speech, and an experiment was made on the discrimination of personalities. As a result, it is seen that the personality is strongly sensitive to the formant frequency shift, especially to the shift of the lower three formants, and it is lost by the uniform shift of 5 percent. On the other hand, the personality is well preserved for the change of formant bandwidth, and is sensitive to the bandwidth change in the fourth‐ and higher‐order for‐mants. In the case of uniform change of the bandwidth, deformation of wider than 5 times or narrower than one‐fifth times completely eliminates the pe
ISSN:8756-6621
DOI:10.1002/ecja.4410700902
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1987
数据来源: WILEY
|
3. |
Ultrasonic cell with IDT's for measurement of liquid acoustic velocity |
|
Electronics and Communications in Japan (Part I: Communications),
Volume 70,
Issue 9,
1987,
Page 22-29
Showko Shiokawa,
Toyosaka Moriizumi,
Tsuneo Kikuchi,
Preview
|
PDF (701KB)
|
|
摘要:
AbstractThis paper proposes a measuring system for the small changes of sound velocity, using the leaky surface wave excited by the interdigital transducer (IDT) on a piezoelectric substrate. The ultrasonic cell was fabricated and an experimental investigation was made. It is shown that the sound velocity change of the order of 2 ppm can be measured. When the propagation surface is loaded by a liquid, the elastic surface wave excited by IDT is converted into the leaky surface wave, which excites the ultrasound into the liquid. By constructing the measuring system utilizing this ultrasound excitation in the water, little spurious wave is excited, and a differential measurement can be successfully performed in the present system. The characteristics of the fabricated ultrasonic cell are examined and the measurement limit and accuracy are discussed. As examples, the change of sound velocity in KCL solution by concentration and some other measurements are discussed, indicating the usefulness of the proposed measuring system.
ISSN:8756-6621
DOI:10.1002/ecja.4410700903
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1987
数据来源: WILEY
|
4. |
Vowel recognition using an LPC deviation model |
|
Electronics and Communications in Japan (Part I: Communications),
Volume 70,
Issue 9,
1987,
Page 30-43
Itsuo Kumazawa,
Taizo Iijimas,
Preview
|
PDF (753KB)
|
|
摘要:
AbstractWhen the linear prediction coefficients (LPC) are used as a feature parameter of speech, a problem in speech recognition is the variation of LPC due to the difference of the speaker and the effect of the preceding or succeeding utterances. Because of this variation, LPC is distributed in a certain region in the LPC space. This paper proposes a method which approximates the region of distribution by a linear manifold model described in [1], and recognizes the category permitting the variation of LPC. The method is applied to the vowel recognition of CV syllable, and its usefulness is verified experimentally. It is shown first by spectrum analysis experiment that the linear manifold model is useful in the approximation of the variation of the vowel spectrum due to the preceding consonant. Second, a recognition experiment is carried out based on the model, and it is shown that the recognition rate can be improved by a simple model of linear manifold which is suited to computation.
ISSN:8756-6621
DOI:10.1002/ecja.4410700904
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1987
数据来源: WILEY
|
5. |
Computational accuracy in determining the sign of an algebraic quantity |
|
Electronics and Communications in Japan (Part I: Communications),
Volume 70,
Issue 9,
1987,
Page 44-49
Kiyonori Yoshida,
Preview
|
PDF (364KB)
|
|
摘要:
AbstractThis paper considers the computational accuracy when performing a numerical computation of an algebraic quantity to determine the sign. For an algebraic function α({di}) of data {di} we introduce the (n,m,l) degree to discuss a method for its estimation. It is shown that when data {di} are given in fixed decimal point form with finite figures, we can correctly determine the sign of the value of α with approximatelymtimes more figure
ISSN:8756-6621
DOI:10.1002/ecja.4410700905
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1987
数据来源: WILEY
|
6. |
A method for reducing the number of product terms in a symbolic reliability expression |
|
Electronics and Communications in Japan (Part I: Communications),
Volume 70,
Issue 9,
1987,
Page 50-58
Youichi Higashiyama,
Hiromu Ariyoshi,
Preview
|
PDF (514KB)
|
|
摘要:
AbstractAmong methods for deriving a symbolic reliability expression between two specified vertices (s,t) in a network, a treeexpansion method proposed herein generates an expression in summation form with fewer product terms and computes it very fast. The tree‐expansion method has a fundamental operation to enumerates–tpaths {μ1, μ2, …, μm} and generates a rooted tree (reference tree) from subpaths of paths μj(J= 1, 2, …,i– 1) having their initial and terminal vertices on a path μiand then derives a symbolic reliability expression directly from its subtree (implicant tree). This method has a disadvantage in that the number of product terms depends heavily on the order of path generation since the reference tree must be generated according to the order of paths obtained by a depth‐first search. In this paper for each vertex in a network we define its weight as the number of edges in a shortest path from the vertex tot. We show that if we generate paths by a depth‐first search to prefer those edges with terminal vertices having small weights and apply the tree‐expansion method according to the generation order, then we can reduce the number of product terms (this is called the modified tree‐expansion method). Also, we show by computer experiments that the modified tree‐expansion method can reduce the number of product te
ISSN:8756-6621
DOI:10.1002/ecja.4410700906
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1987
数据来源: WILEY
|
7. |
Time complexity of gate assignment problem in one‐dimensional array |
|
Electronics and Communications in Japan (Part I: Communications),
Volume 70,
Issue 9,
1987,
Page 59-67
Takashi Fujii,
Tohru Kikuno,
Noriyoshi Yoshidas,
Hideya Horikawa,
Preview
|
PDF (590KB)
|
|
摘要:
AbstractIn the one‐dimensional array approach, a logic circuit is realized by arranging NAND (NOR) gates in a one‐dimensional array and interconnecting them. The horizontal length of the resultant layout of the logic circuit is determined uniquely by the number of given gates, while the vertical length varies depending on the order of assignment of gates. Thus, to minimize the area required, we must find a one‐dimensional assignment of gates so as to minimize the vertical length. This optimization problem (called problem G) has been formulated as a graph problem and has already been proved to be NP‐hard.In this paper, a restricted problem (called the problem R) is introduced anew wherein a connection of each net is restricted to be between two gates. The problem R plays an important role in developing a heuristic algorithm for the problem G. As for the time complexity, we show that the problem R is also
ISSN:8756-6621
DOI:10.1002/ecja.4410700907
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1987
数据来源: WILEY
|
8. |
A rearrangement algorithm for three‐stage switching networks |
|
Electronics and Communications in Japan (Part I: Communications),
Volume 70,
Issue 9,
1987,
Page 68-77
Satoru Ohta,
Hiromi Ueda,
Preview
|
PDF (667KB)
|
|
摘要:
AbstractIn the three‐stage rearrangeable switch‐ing network, an internal blocking can occur. However, such a blocking can be eliminated by rearranging the already set‐up paths, and thus the network can satisfy any connection requirement. To eliminate this internal blocking, a successive rearrangement of more than one already set‐up path is required. The sequence of the paths to be rearranged can clearly be represented by extending the notion of the connection chain introduced by Gotoh. This paper presents a sufficient condition for the Internal blocking to be eliminated by the rearrangement in the extended connection chain, which is a basis for the rearrangement algorithm. It is also verified that the rearrangement algorithm proposed by Gotoh satisfies the sufficient condition. Based on these results, a connection chain is shown in which the internal blocking can be eliminated by the rearrangement. The algorithm to search for the connection chain and to perform the rearrangement is presented. Compared with past methods, in the proposed method, a fewer number of paths is distorted. The usefulness of the method is examined by a sim
ISSN:8756-6621
DOI:10.1002/ecja.4410700908
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1987
数据来源: WILEY
|
9. |
σgi [x i]/m/s/k queues with mixed batch renewal inputs |
|
Electronics and Communications in Japan (Part I: Communications),
Volume 70,
Issue 9,
1987,
Page 78-86
Fumiaki Machihara,
Preview
|
PDF (583KB)
|
|
摘要:
AbstractThis paper presents the analysis for the finite capacity queueing model with two independent mixed batch renewal inputs. The analysis is presented for the steady‐state distribution at an arbitrary time and the steady‐state distribution immediately before the arrival for each type of customers. The results can be applied to the models of various kinds of strategies such as delay‐delay, delay‐loss and loss‐loss. The structure of this paper is as follows. First, a simple numerical calculation procedures is proposed for the distribution at an arbitrary time which, compared with past methods, drastically reduces the computational complexity. Then the relationship is established between the distribution at an arbitrary time and the distribution immediately before the arrival, using the rate conservation principle in the theory of piecewise Markov process. Using this relationship, one distribution can easily be derived from the other. The individual blocking probabilities, which are the most important performance measures, are derived from the distribution immediately before the arrival. Finally, several examples are presented in order to discuss the effects of the interarrival distribution and the batch‐size distribution on the individual blocking pr
ISSN:8756-6621
DOI:10.1002/ecja.4410700909
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1987
数据来源: WILEY
|
10. |
A multiplexing system for digital transmission and process in vehicles |
|
Electronics and Communications in Japan (Part I: Communications),
Volume 70,
Issue 9,
1987,
Page 87-94
Aiguo He,
Akira Kawamatas,
Preview
|
PDF (541KB)
|
|
摘要:
AbstractAt present, the vehicular electronic system is composed of several independent subsystems. By contrast, this paper proposes an integrated system where the same functions are realized in a more reasonable and economical way, utilizing the digital multiplex communication and distributed control. The system aims at optimization of the connections and control of the components, considering only the distances among the subsystems and components, independently of their kinds and functions. To improve the economy, reliability and flexibility, the system is composed of a ring transmission system and a distributed processing system using the microprocessors and optical fibers. This paper discusses the functions and structures of the transmission and processing devices composing the system. The optimum number of transmission and processing devices as well as the processing ability of the system are examined.
ISSN:8756-6621
DOI:10.1002/ecja.4410700910
出版商:Wiley Subscription Services, Inc., A Wiley Company
年代:1987
数据来源: WILEY
|
|