Browsing by Issue Date, starting with "1997-02-28"

Now showing 1 - 20 of 83996

Faithful 1-edge fault tolerant graphs
(1997-02-28)
Wang, SY
;
Hsu, LH
;
Sung, TY
;
資訊工程學系
;
Department of Computer Science
A graph G* is 1-edge fault tolerant with respect to a graph G, denoted by 1-EFT(G), if any graph obtained by removing an edge from G* contains G. A 1-EFT(G) graph is said to be optimal if it contains the minimum number of edges among all 1-EFT(G) graphs. Let G(i)* be 1-EFT(G(i)) for i = 1,2. It can be easily verified that the cartesian product graph G(1)* x G(2)* is 1-edge fault tolerant with respect to the cartesian product graph G(1) x G(2). However, G(1)* x G(2)* may contain too many edges; hence it may nor be optimal for many cases. In this paper, we introduce the concept of faithful graph with respect to a given graph, which is proved to be 1-edge fault tolerant. Based on this concept, we present a construction method of the 1-EFT graph for the cartesian product of several graphs. Applying this construction scheme, we can obtain optimal 1-edge fault tolerant graphs with respect to n-dimensional tori C(m(1), m(2),...,m(n)), where m(i) greater than or equal to 4 are even integers for all 1 less than or equal to i less than or equal to n. (C) 1997 Elsevier Science B.V.
Faithful 1-edge fault tolerant graphs
(1997-02-28)
Wang, SY
;
Hsu, LH
;
Sung, TY
;
資訊工程學系
;
Department of Computer Science
A graph G* is 1-edge fault tolerant with respect to a graph G, denoted by 1-EFT(G), if any graph obtained by removing an edge from G* contains G. A 1-EFT(G) graph is said to be optimal if it contains the minimum number of edges among all 1-EFT(G) graphs. Let G(i)* be 1-EFT(G(i)) for i = 1,2. It can be easily verified that the cartesian product graph G(1)* x G(2)* is 1-edge fault tolerant with respect to the cartesian product graph G(1) x G(2). However, G(1)* x G(2)* may contain too many edges; hence it may nor be optimal for many cases. In this paper, we introduce the concept of faithful graph with respect to a given graph, which is proved to be 1-edge fault tolerant. Based on this concept, we present a construction method of the 1-EFT graph for the cartesian product of several graphs. Applying this construction scheme, we can obtain optimal 1-edge fault tolerant graphs with respect to n-dimensional tori C(m(1), m(2),...,m(n)), where m(i) greater than or equal to 4 are even integers for all 1 less than or equal to i less than or equal to n. (C) 1997 Elsevier Science B.V.
A novel shallow trench isolation technique
(1997-03-01)
Cheng, JY
;
Lei, TF
;
Chao, TS
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
Shallow-trench isolation processes which involve refilling using deposition of oxide and polysilicon were investigated. Our results show that the oxide-filled shallow-trench isolation technology based on chemical-mechanical polishing (CMP) is difficult to control and results in poor uniformity. Use of this technology also involves in the dishing effect in wide field regions. However, shallow-trench isolation technology using a masking nitride layer, polysilicon refill, the CMP process with high etch selectivity, and local oxidation of polysilicon is easily implemented and absolutely field oxide encroachment (bird's beak) free. Although the CMP process results in the dishing effect in wide field regions, the local oxidation of polysilicon can reduce the amount of dishing. The polysilicon-filled shallow-trench isolation process can also achieve excellent uniformity across 6-inch diameter silicon wafers due to the high etching selectivity of polysilicon to chemical-vapor-deposited (CVD) oxide and SiN. Moreover, the n(+)/p junction leakage current of polysilicon-filled shallow trenches is comparable to that of oxide-filled shallow trenches. This simple and easily controllable process is a very promising candidate for shallow trench isolation.
The generalized Borel-Weil theorem and cohomology of G/(P,P)
(1997-03-01)
Chuah, MK
;
應用數學系
;
Department of Applied Mathematics
Let G be a complex semi-simple Lie group. The generalized Borel-Weil theorem computes the cohomology of a compact homogeneous G-space with coefficient in homogeneous line bundle. In this paper, we give an analogous statement via Dolbeault cohomology of the non-compact space G/(P,P) for P subset of G a parabolic subgroup, and (P,P) the commutator subgroup.
Planar constrained terminals over-the-cell router
(1997-03-01)
Shew, PW
;
Hsiao, PY
;
資訊工程學系
;
Department of Computer Science
The authors present a new routing model for over-the-cell channel routing. A graph theoretical algorithm is then proposed to solve the new problem, The algorithm has a complexity of O(nk(2)), where n is the number of nets and k is the number of columns in the channel. It achieved a routing area reduction of 71.5% for the PRIMARY 1 benchmark example from MCNC, using three-layer over-the-cell routing. To resolve a sub-problem, the authors also present an O(m nu) algorithm to find the maximum weight independent chord set in a circle graph with m chords incident to nu vertices, where two chords may share a common vertex.
Application of multiple criteria decision making for network improvement
(1997-03-01)
Tzeng, GH
;
Tsaur, SH
;
運輸與物流管理系註：原交通所+運管所
;
Department of Transportation and Logistics Management
In this paper we propose application of multiple criteria decision making to problems of a metropolitan network improvement plan. Initially, a bilevel multiple objective network design model is considered in two objectives which are minimal government budget and minimal total travel time of road users. We seek feasible improvement alternatives among those bottleneck links in an existing road network structure and travel demand. We present an effective heuristic algorithm to obtain noninferior solutions; then ELECTRE III multiple criteria decision making and group decision making are used to evaluate and to select a compromise solution among those noninferior solutions. From the design phase in multiple criteria decision making, multiple objective mathematical programming is used to formulate a continuous network design model. However, from the phase of evaluation, multiple criteria decision making to solve the discrete network design problem. The network of metropolitan Taipei is taken as an example to illustrate the operation of this model.
A practical design for a robust fault detection and isolation system
(1997-03-01)
Hwang, DS
;
Chang, SK
;
Hsu, PL
;
電控工程研究所
;
Institute of Electrical and Control Engineering
Based on a development for the unknown input observer (UIO) design, this paper presents a straightforward method of designing a robust fault detection and isolation (FDI) system for uncertain systems with unknown inputs. By exploiting the properties of a nominal model structure, we form a basic sec of residuals which is independent of the unknown input but is dependent on the measurements. Then, with this basic residual set, we append a dynamic weighting matrix to the residuals to increase the design freedom for meeting the required detection and isolation performance. Additionally, a generalized residual generator scheme (GRGS) is proposed to synthesize the parity relations for the isolation of faulty sensors in an uncertain system. The proposed strategy is verified through simulation studies performed on the control of a vertical take-off and landing (VTOL) aircraft in the vertical plane.
Heuristic algorithm for the design of facilities layout and AGV routes in tandem AGV systems
(1997-03-01)
Hsieh, LF
;
Sha, DY
;
交大名義發表
;
工業工程與管理學系
;
National Chiao Tung University
Considering the necessity for the integrated design of facilities layout and material handling systems, this paper develops a heuristic algorithm to solve the facilities layout and AGV routes problems simultaneously in a tandem automated guided vehicle (AGV) system. A tandem AGV system is composed of several nonoverlapping closed loops; each loop is traversed by a vehicle. Based on this system in which bi-directional vehicles are used, distance and material now between pickup/deposit stations are considered simultaneously during the design of the facilities layout. In the meantime, several transit stations would handle the transportation between loops. As its subject, this paper proposes a heuristic algorithm to implement the facilities layout within each loop and the route of each vehicle in order to reduce the material handling cost It is proved that the heuristic algorithm is polynomial time complexity with O(n(4)).
A transputer-based adaptive speed controller for AC induction motor drives with load torque estimation
(1997-03-01)
Tsai, MF
;
Tzou, YY
;
交大名義發表
;
電控工程研究所
;
National Chiao Tung University
In this paper, we design and implement an adaptive speed controller that can estimate load torque for ac induction motor drives employing a transputer-based parallel processing technique, The adaptive speed controller, which precedes the held-oriented control loop, consists of a two-degree-of-freedom controller and a feedforward load-torque compensator, The two-degree-of-freedom controller is designed by a pole-placement technique with polynomial manipulations. Its parameters are adjusted adaptively in terms of estimated model parameters, Estimating the model parameters entails a second-order least-squares estimator with constant trace to avoid estimator windup. The design of the feedforward compensator is based on an estimated load-torque model, Estimating the load torque entails a first-order least-squares estimator with variable forgetting factor and covariance resetting, the purposes of which are to detect any slow or sudden changes of torque disturbance, respectively. The resulting adaptive controller is implemented in parallel by IMS T800-20 transputers, Experimental results demonstrate the robustness of the proposed control method in contending with varying load and torque disturbance.
Design of general structured observers for linear systems with unknown inputs
(1997-03-01)
Chang, SK
;
You, WT
;
Hsu, PL
;
生醫工程研究所
;
電控工程研究所
Using the framework of the general structured (GS) observer, we present a straightforward procedure for designing an unknown input observer (UIO) for a linear system subject to unknown inputs ol uncertain disturbances. The set of all GS observers insensitive to unknown inputs is derived in this paper. Moreover, an extension of the UIO, called the extended UIO, is developed to estimate both the system state and the unknown input simultaneously. We show the existence conditions of a stable UIO are the same as those of a stable left inverse system. III addition, well-conditioned designs for both the state and unknown input estimations ale also explored. Conditions of transmission zeros reveal that to achieve a stable UIO, the uncertain system should be minimum-phase. To overcome this restriction, we adopt a two-delay output stabilized method to design the stabilized UIO without implementing extra sensors. Experimental results for a DC servo motor system demonstrate the applicability of the proposed methodologies. Copyright (C) 1997 Published by Elsevier Science Ltd.
Design of general structured observers for linear systems with unknown inputs
(1997-03-01)
Chang, SK
;
You, WT
;
Hsu, PL
;
電控工程研究所
;
Institute of Electrical and Control Engineering
Using the framework of the general structured (GS) observer, we present a straightforward procedure for designing an unknown input observer (UIO) for a linear system subject to unknown inputs ol uncertain disturbances. The set of all GS observers insensitive to unknown inputs is derived in this paper. Moreover, an extension of the UIO, called the extended UIO, is developed to estimate both the system state and the unknown input simultaneously. We show the existence conditions of a stable UIO are the same as those of a stable left inverse system. III addition, well-conditioned designs for both the state and unknown input estimations ale also explored. Conditions of transmission zeros reveal that to achieve a stable UIO, the uncertain system should be minimum-phase. To overcome this restriction, we adopt a two-delay output stabilized method to design the stabilized UIO without implementing extra sensors. Experimental results for a DC servo motor system demonstrate the applicability of the proposed methodologies. Copyright (C) 1997 Published by Elsevier Science Ltd.
Unsupervised query-based learning of neural networks using selective-attention and self-regulation
(1997-03-01)
Chang, RI
;
Hsiao, PY
;
交大名義發表
;
資訊工程學系
;
National Chiao Tung University
Query-based learning (QBL) has been introduced for training supervised network model with additional queried samples, Experiments demonstrated that the classification accuracy is further increased. Although QBL has been successfully applied to supervised neural networks, it is not suitable for unsupervised learning models without external supervisors. In this paper, an unsupervised QBL (UQBL) algorithm using selective-attention and self-regulation is proposed. Applying the selective-attention, we can ask the network to respond to its goal-directed behavior with self-focus. Since there is no supervisor to verify the self-focus, a compromise is then made to environment-focus with self-regulation. In this paper, we introduce UQBL1 and UQBL2 as two versions of UQBL; both of them can provide fast convergence. Our experiments indicate that the proposed methods are more insensitive to network initialization. They have better : generalization performance and can be a significant reduction in their training size.
Bifurcations of symmetric periodic orbits near equilibrium in reversible systems
(1997-03-01)
Shih, CW
;
交大名義發表
;
應用數學系
;
National Chiao Tung University
;
Department of Applied Mathematics
Consider a family of reversible systems (x) over dot = f(x, mu) with the origin being an equilibrium for each mu. Suppose D(x)f(0, 0) has only purely imaginary eigenvalues +/-iw(1),..., +/-iw(k). We investigate the typical bifurcations of symmetric periodic solutions near the origin. A suitable complex basis is chosen so that D(x)f(0, 0) and the involution are in respective simple form. Incorporated with putting f into normal form, a modified version of Lyapunov-Schmidt reduction can be applied to obtain the reduced bifurcation equations. We then focus on the cases in resonance, that is, w(j) = n(j)w(0), where w(0) is a nonzero real number and n(j) is an integer for each j. Some codimension-two bifurcations are illustrated for the system in non-semisimple resonance with n(j) = 1, 2. A few codimension-one cases are also given for comparison with earlier works by other researchers.
A general mathematical model for gears cut by CNC hobbing machines
(1997-03-01)
Chang, SL
;
Tsay, CB
;
Nagata, S
;
機械工程學系
;
Department of Mechanical Engineering
A hobbing machine's cutting mechanism is a mechanism with multi-degree of freedom during the cutting process. In this paper, we propose a general gear mathematical model simulating the generation process of a 6-axis CNC hobbing machine based on the cutting mechanism of CNC hobbing machine and worm-type hob cutter. The proposed gear mathematical model can be applied to simulate different types of gear cutting. Some examples are included to verify the mathematical model. Also, a novel type of gear named ''Helipoid'' which can be used in crossed axes transmission is proposed. The proposed general gear mathematical model can facilitate a more thorough understanding of generation processes and toward the development of novel types of gears.
Elementary excitations in a carbon nanotube
(1997-03-01)
Lin, MF
;
Chuu, DS
;
交大名義發表
;
電子物理學系
;
National Chiao Tung University
A carbon nanotube exhibits the decoupled pi plasmons of different angular momenta (L's). The magnetic Aux is expected to be useful in identifying them. All the L-decoupled plasmons have strong dispersion relations with momentum (q). The L not equal 0 plasmons belong to optical plasmons, while the L = 0 plasmon could not exist at q --> 0. The former, respectively, behave as the plasmons in three- and two-dimensional electron gases at small and large q's. The plasmon frequencies would be affected by radius, chiral angle, and intertube interaction. To obtain the actual plasmon frequencies of a multi-shell nanotube, the intertube interactions from the neighboring shells have to be taken into account. The calculated results are consistent with the experimental measurements. A detailed comparison between a carbon nanotube and a graphite layer is also made. These two related systems may exhibit similar pi plasmons, when L = 0 in a nanotube. Furthermore, the chirality effect of the former corresponds to the anisotropy effect of the latter. The predicted q-, L-, and geometry-dependence of the pi plasmons needs further experimental verifications.
The effect of grain boundary impedance on the power loss of Mn-Zn ferrites over 1 MHz
(1997-03-01)
Tung, MJ
;
Tseng, TY
;
Tsay, MJ
;
Chang, WC
;
電子工程學系及電子研究所
Recently, the operating frequency of switching mode power supplies has been increased from 25 kHz to several MHz to enhance their efficiency in a smaller size. Although, it was reported that Mn-Zn ferrites can also be used at 1 MHz, only few papers have discussed the possibility of using Mn-Zn ferrites at the frequency higher than 1 MHz. Our early paper reported that the core loss of low loss grade Mn-Zn ferrites is strongly dependent upon the capacitance of grain boundaries of sintered cores. In frequency region between 1 MHz to 3 MHz, the dielectric loss dominates the core loss, while eddy current loss is the major factor of the core loss in 3 MHz to 10 MHz. It is suggested that decreasing the grain boundary capacitance at middle high frequency or increasing the grain resistivity at high frequency region are the appropriate ways for Mn-Zn ferrites to be used at the frequency between 1 MHz and 10 MHz. This paper investigated the core loss of Mn-Zn ferrites in the frequency range between 1 MHz and 10 MHz in order to improve the properties of the Mn-Zn ferrite cores for their potential applications in this frequency region. The powders containing Fe2O3, MnO and ZnO ina molar ratio of 54.2:7.3:8.5 with the addition of CaO and SiO2 from 150 ppm to 450 ppm were prepared by a conventional ceramic process. Samples were sintered at 1150 degrees C to 1250 degrees C for 2 hrs then annealed at various condition. The results show that the sintering temperature and annealing condition altered the grain boundary impedance. For example, the grain boundary capacitance changes from 13 pF to 7 pF, meanwhile, the dielectric loss factor is lowered from 0.28 x 10(-9) to 0.17 x 10(-9), and consequently the power loss decreases from 253 KW/m(3) to 145 KW/m(3) at 2.5 MHz.
法國這個，不，這一隻(遊法印象)
(1997-03-01)
管管
On liquid-phase deposition of silicon dioxide by boric acid addition
(1997-03-01)
Chang, PH
;
Huang, CT
;
Shie, JS
;
材料科學與工程學系
;
光電工程學系
The current status of liquid-phase deposition (LPD) of SiO2 by adding boric acid (H3BO3) to a hydrofluosilicic acid (H2SiF6) is reviewed and compared with some new results from the authors' laboratory. Large discrepancies exist in the literature concerning the effects of Various processing parameters on deposition rate. We have shown that much confusion arises from the misconception of using an SiO2 additive to ''saturate'' and usng water to ''dilute'' the growth solution. In this paper the role of H2O as a reagent and the detrimental effect of an SiO2 additive as seeds for depleting the nutrient in the growth solution if added after the H2O addition are emphasized. Despite the variations between different investigators, the following characteristics of the LPD process are in general agreement: (i) the deposition process is surface controlled, (ii) the deposition rate depends approximately linearly on temperature, (iii) the deposition rate is constant (independent of time) over a wide range of experimental conditions, and (iv) the deposition rate increases with increasing boric acid concentration.
Ionization of oriented elliptic Rydberg states by half-cycle pulses
(1997-03-01)
Jiang, TF
;
Lin, CD
;
物理研究所
;
Institute of Physics
We calculate the ionization of elliptic Rydberg states by half-cycle unipolar electromagnetic-field pulses and examine the dependence of ionization probabilities on the eccentricity and the orientation of the ellipse with respect to the field direction of the pulse. It is suggested that the predicted large geometrical dependences be tested experimentally.
Design of a mathematical expression understanding system
(1997-03-01)
Lee, HJ
;
Wang, JS
;
交大名義發表
;
資訊工程學系
;
National Chiao Tung University
A scientific document usually consists of text and mathematical expressions. In this paper, we present a system for segmenting and understanding text and mathematical expressions in a document, The system can be divided into six stages: page segmentation and labeling, character segmentation, feature extraction, character recognition, expression formation, and error correction and expression extraction. After we extract all text lines in a document, we separate all symbols in each text line and calculate direction-feature vectors and aspect ratios for those symbols. Then, a nearest-neighbor algorithm recognizes characters. In the expression formation stage, we build a symbol relation tree for each text line that represents the relationships among the symbols in the text line. Each text line is decomposed into a collection of primitive tokens: operands, operators and separators. Heuristic rules based on these primitive tokens are used to correct text recognition errors. Finally, we extract all mathematical expressions according to basic expression forms. Several pages of documents were scanned to test the method. All mathematical expressions are understood. In the expressions generated, a few symbols are misrecognized. The average recognition rate was 96.16%. (C) 1997 Elsevier Science B.V.