Browsing by Issue Date, starting with "1994-02-01"

Now showing 1 - 20 of 88454

PERFORMANCE ANALYSIS OF AN OCR SYSTEM VIA AN ARTIFICIAL HANDWRITTEN CHINESE CHARACTER GENERATOR
(1994-02-01)
TUNG, CH
;
CHEN, YJ
;
LEE, HJ
;
資訊工程學系
;
Department of Computer Science
A handwritten Chinese character generator that generates handwritten Chinese characters with different variations is proposed and used to evaluate the performance of an OCR system. The character generator first generates radicals of a character stroke by stroke and then combines the generated radicals to form a line-vector character. Next, the line-vector character is thickened to obtain a character image. Characters generated by this method will have great variance in shape but will still satisfy the structural constraints. The generated characters are then used to perform two types of evaluations. First, the stability of stroke extractors in terms of stroke number is evaluated. Two stroke extractors, one based on thinning and the other on vectorization, are analyzed. The vectorization-based stroke extractor we propose operates directly on the run length codes of line segments. From experimental results, we find that the thinning-based method is more time-consuming but more stable than the vectorization-based method. Second, the peak performance of a matching module is evaluated and the recognition error caused by stroke extractors is identified.
CONCEPTUAL SURVEY OF PHASE MODULATION BASED WAVELENGTH REGISTERED NETWORKS
(1994-02-01)
KAO, MS
;
LIN, TC
;
電信工程研究所
;
電信研究中心
;
Institute of Communications Engineering
The implementation of a dense wavelength-division multiplexing (WDM) based wide-area network inevitably involves a large number of semiconductor lasers requiring particular temperature and driving current control. The concept of a novel multiwavelength network that uses only one laser source to generate all the carriers needed for the entire network is investigated. The central idea is to use a stable reference laser to generate two sets of optical carriers as the 'carrier supplies' for all network nodes: one set for signal transmission and the other for signal reception. A promising feature of the proposed network is the wavelength registration, that is, all carrier wavelengths are unambiguously identified. This special feature is of importance for a standard wide-area multiwavelength network. Optical phase modulation is extensively employed in the carrier generation as well as in information delivery. A 40-wavelength network is considered with each wavelength carrying 20 phase-modulated microwave subcarriers, a total throughput of 160 Gbit/s per single-mode fibre is estimated.
A NEW OBJECT-ORIENTED METHOD INTEGRATING JACKSON STRUCTURED PROGRAMMING METHOD
(1994-02-01)
CHEN, JY
;
CHEN, YTT
;
資訊工程學系
;
Department of Computer Science
This article presents a new object-oriented method which integrates the Jackson Structured Programming method with an object-oriented approach. The new method analyzes the user's services, identifies attributes from data structure diagrams, and identifies operations from program structure diagrams. Also, this method defers the identification of class relationships until the attributes and operations have been identified, so as to prevent premature decisions. In addition, the concept of ''leading object'' is used to decrease the number of actors and agents and to increase the number of servers in a system. This may enhance reusability of the objects.
MAGNETIC-BEHAVIOR IN PR-CONTAINING TL-BASED AND PB-BASED CUPRATES
(1994-02-01)
KU, HC
;
LAI, CC
;
SHIEH, JH
;
LIOU, JW
;
WU, CY
Magnetic ordering among Pr ions occurs at a relatively high temperature of 17 K in PrBa2Cu3O7, but is absent in Pr2CuO4 and Bi2Sr2PrCu2O8 above 1.6 K. This work extends the study of magnetic properties to other Pr-containing cuprates including TlBa2PrCu2O7, (Pb.6Cu.4)Sr2PrCu2O7 and Pb2Sr2PrCu3O8. Low temperature calorimetric and SQUID measurements show broad peaks in C and dchi/dT in all cases. The onset temperatures are near 10 K, while the maxima in C as well as minima in dchi/dT prevail at about 8 to 5 K. The heat capacity data provides additional information concerning the entropy associated with the magnetic anomalies and large contributions seemingly linear with temperature.
REINFORCEMENT STRUCTURE PARAMETER LEARNING FOR NEURAL-NETWORK-BASED FUZZY-LOGIC CONTROL-SYSTEMS
(1994-02-01)
LIN, CT
;
LEE, CSG
;
電控工程研究所
;
Institute of Electrical and Control Engineering
This paper proposes a reinforcement neural-network-based fuzzy logic control system (RNN-FLCS) for solving various reinforcement learning problems. The proposed RNN-FLCS is constructed by integrating two neural-network-based fuzzy logic controllers (NN-FLC's), each of which is a connectionist model with a feedforward multilayered, network developed for the realization of a fuzzy logic controller. One NN-FLC performs as a fuzzy predictor, and the other as a fuzzy controller. Using the temporal difference prediction method, the fuzzy predictor can predict the external reinforcement signal and provide a more informative internal reinforcement signal to the fuzzy controller. The fuzzy controller performs a stochastic exploratory algorithm to adapt itself according to the internal reinforcement signal. During the learning process, both structure learning and parameter learning are performed simultaneously in the two NN-FLC's using the fuzzy similarity measure. The proposed RNN-FLCS can construct a fuzzy logic control and decision-making system automatically and dynamically through a reward/penalty signal (i.e., a ''good'' or ''bad'' signal) or through very simple fuzzy information feedback such as ''high,'' ''too high,'' ''low,'' and ''too low.'' The proposed RNN-FLCS is best applied to the learning environment, where obtaining exact training data is expensive. The proposed RNN-FLCS also preserves the advantages of the original NN-FLC, such as the ability to find proper network structure and parameters simultaneously and dynamically and to avoid the rule-matching time of the inference engine in the traditional fuzzy logic systems. Computer simulations were conducted to illustrate the performance and applicability of the proposed RNN-FLCS.
ANALYSIS OF BILATERAL LATCH-UP TRIGGERING IN VLSI CIRCUITS
(1994-02-01)
HUANG, HS
;
CHANG, CY
;
HSU, CC
;
CHEN, KL
;
LIN, JK
4-POSITION SYNTHESIS FOR SPATIAL MECHANISMS WITH 2 INDEPENDENT LOOPS
(1994-02-01)
CHIANG, CH
;
YANG, YN
;
CHIENG, WH
;
HOELTZEL, DA
;
機械工程學系
During the practical implementation of four-position guidance synthesis for the spatial mechanisms, many mathematical models for rigid body guidance design are found to be too complicated to solve especially for up to four positions. Our objectives are to uncover simple and analytical ways to synthesize the single degree-of-freedom, two independent loop spatial mechanisms which are capable of four precision positions. Furthermore, prescribed timing is simultaneously considered for some spatial mechanisms.
TRANSIENT ANALYSIS OF SUBMICRON CMOS LATCHUP WITH A PHYSICAL CRITERION
(1994-02-01)
KER, MD
;
WU, CY
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
A new physical criterion and an analytical timing model for transient latchup in a p-n-p-n structure are developed and verified. In the new physical transient latchup criterion, two new parameters called the large-signal transient current gains of the parasitic vertical and lateral BJTs in the p-n-p-n structure are defined. If their product rather than the forward beta-gain product becomes larger than unity right after triggering and remains larger than unity, the latchup occurs. With the piecewise-linearized device currents of the parasitic BJTs and the averaged junction depletion and diffusion capacitances in the p-n-p-n structure, the large-signal transient current gains are derived as functions of time and device parameters. Moreover, the dynamic behaviors and the associated timing of transient latchup can also be fully characterized. Model calculation results using the developed criterion and timing model agree very well with both SPICE simulation and experimental results. Thus the developed criterion and timing model can be applied to transient latchup analysis and prediction in CMOS ICs.
VIRTUAL KNOT TECHNIQUE FOR CURVE-FITTING OF RAPIDLY VARYING DATA
(1994-02-01)
HSIEH, HC
;
CHANG, WT
;
電信工程研究所
;
Institute of Communications Engineering
Curve fitting with piecewise splines for rapidly varying data has been a difficult problem. Often, the curve exhibits unwanted ''wiggles'' around those data points whose locations change rapidly in comparison to their neighboring data points. Many methods were proposed to solve this problem using non-homogeneous tension spline or fairing process. In this paper, a new approach using the concept of virtual knots is presented. Since the occurrence of the problem is due to the inconsistency between the parametric spans of the knots and the geometric spans of the data points, a number of virtual knots are inserted into the original knot sequence such that the parametric spans are made more consistent with the geometric spans of the data. With this parametric adjustment, a regularization process is used to find the desired curve. Experimental results show that this technique is efficient to adjust the fairness of the desired curve. This technique can be applied to different spline spaces. In this paper, the usages of this technique in linear B-spline and cubic B-spline spaces are demonstrated.
SOME GOOD MULTIPLIERS FOR RANDOM NUMBER GENERATOR FOR 16-BIT MICROCOMPUTERS
(1994-02-01)
CHANG, PL
;
HWANG, SN
;
KAO, C
;
管理科學系
;
Department of Management Science
In searching for statistically sound multipliers for the prime modulus multiplicative congruential random number generator for 16-bit microcomputers, an exhaustive analysis is conducted over the 10,912 primitive roots of the modulus 2(15)-19=32,749. The result shows that there are only 227 good multipliers among the 10,912 candidate multipliers. Here a multiplier is said to be good if its corresponding generator has good statistical properties and computational efficiency.
GROWTH OF PULSED-LASER DEPOSITED YBCO THIN-FILMS ON YSZ AND YSZ/SI SUBSTRATES
(1994-02-01)
JUANG, JY
;
KAO, RT
;
HSIEH, MC
;
LI, MH
;
CHU, ML
The growth mechanism of superconducting YBCO thin films deposited in - situ on both single crystalline YSZ and YSZ-buffered Si substrates as revealed by scanning tunneling microscopy (STM) are found to be very different. The results are attributed to the initial surface condition of the substrates.
GENERALIZED ON-SHELL WARD IDENTITIES IN STRING THEORY
(1994-02-01)
LEE, JC
;
交大名義發表
;
物理研究所
;
National Chiao Tung University
;
Institute of Physics
It is demonstrated that an infinite set of string-tree level on-shell Ward identities, which are valid to all sigma-model loop orders, can be systematically constructed without referring to the string field theory. As examples, bosonic massive scattering amplitudes are calculated explicitly up to the second massive excited states. Ward identities satisfied by these amplitudes are derived by using zero-norm states in the spectrum. In particular, the inter-particle Ward identity generated by the D2XD2, zero-norm state at the second massive level is demonstrated. The four physical propagating states of this mass level are then shown to form a large gauge multiplet. This result justifies our previous consideration on higher inter-spin symmetry from the generalized worldsheet sigma-model point of view.
COMPARATIVE-STUDIES OF GD-ORDERING IN VARIOUS CUPRATE SYSTEMS
(1994-02-01)
HO, JC
;
WU, CY
;
LAI, CC
;
SHIEH, JH
;
KU, HC
Gd-ordering near 2.2 K in both superconducting and nonsuperconducting GdBa2Cu3O7-y has been well documented. This report summarizes our results from comparative studies on other Gd-containing cuprates (with well-defined Neel temperatures in parentheses) including Bi2Sr2GdCu2O8 (1.6 K), TlBa2GdCu2O7 (2.2 K), TlSr2GdCu2O7 (2.4 K), (Pb.5Cu.5)Sr2GdCu2O7 (2.3 K), Pb2Sr2GdCu3O8 (2.3 K), Gd2CuO4 (6.6 K) and (Gd1.85Ce0.15)CuO4 (5.3 K). The Neel temperature values for the Bi-, Tl-, and Pb-based cuprates remain close to that for PrBa2Cu3O7-y having similar layered structures but somewhat different Gd-Gd distance along the c-axis, whereas the T'(214) compounds become ordered at higher temperatures. Also to be noted is the general occurrence of a broad shoulder below the heat capacity peak, which needs to be taken into account for obtaining the expected magnetic entropy DELTAS = Rln8 associated with the ordering.
RULE-BASE STRUCTURE IDENTIFICATION IN AN ADAPTIVE-NETWORK-BASED FUZZY INFERENCE SYSTEM
(1994-02-01)
SUN, CT
;
交大名義發表
;
資訊工程學系
;
National Chiao Tung University
;
Department of Computer Science
Fuzzy rule-base modeling is the task of identifying the structure and the parameters of a fuzzy IF-THEN rule base so that a desired input/output mapping is achieved. Recently, using adaptive networks to fine-tune membership functions in a fuzzy rule base has received more and more attention. In this paper we summarize Jang's architecture of employing an adaptive network and the Kalman filtering algorithm to identify the system parameters. Given a surface structure, the adaptively adjusted inference system performs well on a number of interpolation problems. We generalize Jang's basic model so that it can be used to solve classification problems by employing parameterized t-norms. We also enhance the model to include weights of importance so that feature selection becomes a component of the modeling scheme. Next, we discuss two ways of identifying system structures based on Jang's architecture. For the top-down approach, we summarize several ways of partitioning the feature space and propose a method of using clustering objective functions to evaluate possible partitions. We analyze the overall learning and operation complexity. In particular, we pinpoint the dilemma between two desired properties: modeling accuracy and pattern matching efficiency. Based on the analysis, we suggest a bottom-up approach of using rule organization to meet the conflicting requirements. We introduce a data structure, called a fuzzy binary boxtree, to organize rules so that the rule base can be matched against input signals with logarithmic efficiency. To preserve the advantage of parallel processing assumed in fuzzy rule-based inference systems, we give a parallel algorithm for pattern matching with a linear speedup. Moreover, as we consider the communication and storage cost of an interpolation model, it is important to extract the essential components of the modeled system and use the rest as a backup. We propose a rule combination mechanism to build a simplified version of the original rule base according to a given focus set. This scheme can be used in various situations of pattern representation or data compression, such as in image coding or in hierarchical pattern recognition.
SENSITIVITY OF FAILURE-DETECTION USING GENERALIZED OBSERVERS - COMMENTS
(1994-02-01)
CHANG, SK
;
HSU, PL
;
電子工程學系及電子研究所
;
電控工程研究所
;
Department of Electronics Engineering and Institute of Electronics
By considering the sensitivity-robustness problem in analytical redundancy-based fault detection schemes, Manquez, H. J. and C. P. Diduch (1992). Sensitivity of failure detection using generalized observers. Automatica, 28, 837-840, proposed a new General Observer to overcome the limits on achievable sensitivity imposed by the structure of classical full-order observers and partial state observers. It is shown here that this General Observer can be found to be equivalent to the partial state observer using the generalized Bezout identity. Two corrected representations of equations (20) and (23) in Manquez and Diduch's paper are also provided.
FLOW CALCULATION IN A LOOP-SCAVENGED TWO-STROKE MOTORED ENGINE
(1994-02-01)
Tsui, Y. -Y.
;
Cheng, H. -P.
;
機械工程學系
;
Department of Mechanical Engineering
A multidimensional calculation method is used to investigate the flow in a motored two-stroke engine. The governing equations are written in a moving-coordinate system such that the grid can move with the piston. Grid lines are added into or deleted from the computational domain, depending on opening or closure of the ports. The EPISO algorithm is modified and adopted as the solution procedure. Calculations are performed on an engine of loop-scavenged type. Details of the gas exchange process and the flow structure in the cylinder are shown. The effects of the engine speed, inlet discharge coefficient and the angle of boost port are examined.
MICRO-PIRANI VACUUM GAUGE
(1994-02-01)
WENG, PK
;
SHIE, JS
;
交大名義發表
;
光電工程學系
;
National Chiao Tung University
A Pirani vacuum sensor has been fabricated by the silicon micromachining technique. A square glass membrane was formed on (100) silicon substrate with a platinum-film resistor coated. The membrane is suspended by its four leads extended to the corners of an etched cavity. This structure can provide both low thermal loss through leads to the substrate and large active area for gaseous heat conduction. It thus can be used as a highly sensitive vacuum sensor. The fabricated sensor has shown a linear response of pressure from 8X10(-5) to 6 Torr with constant-temperature operation. It is found that the low-pressure limit of the vacuum sensor was caused by the noise of the instrument used in the experiment. A new terminology called ''noise equivalent pressure'' thus is definable due to this finding. The physical limit of the noise equivalent pressure is analyzed. Optimization of the device structure and the noise equivalent pressure are also discussed in detail. A new method of ambient-temperature compensation is also proposed and analyzed here.
A METHODOLOGY FOR DERIVING AN ENTITY-RELATIONSHIP MODEL-BASED ON A DATA-FLOW DIAGRAM
(1994-02-01)
KUO, FY
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
This article descirbes an object-oriented methodology for deriving an entity-relationship (ER) model from requirements specified in a data flow diagram (DFD). The methodolgy is top down. It begins with an analysis of the objects described in the DFD to produce an object model. Modeling objects instead of individual data items reduces the number of data elements with which the analyst must be concerned initially. Next, information about data synonyms and interdependency is considered to refine the object model. Guidelines for removing redundant, overlapping descriptions of objects are also proposed and the object model is transformed into an ER model by applying a set of abstraction heuristics. The methodology integrates the DFD-based structured analysis methodology and the ER model so that process and data requirements can be analyzed simultaneously. It enables the system developer to understand the relationships between the processes (embedded in the DFD diagram) and data (described in the ER model) at the early stage of system analysis. Moreover, changes of requirements, either in process or in data, can be correlated easily to foster a quality design of the final system. This methodology can facilitate application development based on an existing data base or on an entirely new data base.
A CHARACTERIZATION TECHNIQUE FOR THE DEGRADATION CHARACTERISTICS OF TI/SI SCHOTTKY-BARRIER DIODES AND OHMIC CONTACTS AFTER THERMAL SILICIDATION
(1994-02-01)
LOU, YS
;
WU, CY
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
The effects of thermal silicidation on both the nonideal current-voltage characteristics of the Schottky-banier diodes and the degradation properties of the ohmic contacts have been self-consistently characterized by using a simple interfacial-layer theory. The abruptly changed nonideality of the measured I-V characteristics as TiSi2 is formed can be interpreted as the variations of the interface properties, such as the thermal-equlibrium barrier height, the interfacial-layer capacitance, and the density distribution of the interface state. Moreover, different work functions for Ti and TiSi2 and the change of the dominant type of the interface states and their density distributions are shown to be responsible for the fluctuations of the thermal-equilibrium barrier height after thermal silicidation. Furthermore, the interface parameters extracted from the fabricated Schottky-barrier diodes are used to calculate and compare the simultaneously-processed specific contact resistivities, and satisfactory agreement is obtained. Thus, the silicidation-related degradation characteristics of the ohmic contacts can be directly related to the variations of the interface properties at the Ti/Si interface.
DYNAMIC STIFFNESS MATRICES FOR 2 CIRCULAR FOUNDATIONS
(1994-02-01)
LIOU, GS
;
交大名義發表
;
土木工程學系
;
National Chiao Tung University
;
Department of Civil Engineering
A systematic procedure is presented for generating dynamic stiffness matrices for two independent circular foundations on an elastic half-space medium. With the technique reported in References 1-3, the analytic solution of three-dimensional (3D) wave equations satisfying the prescribed traction due to the vibration of one circular foundation can be found. Since there are two analytic solutions for two prescribed tractions due to the vibrations of two circular foundations, the principle of superposition must be used to obtain the total solution. The interaction stresses (prescribed tractions) are assumed to be piecewise linear in the r-directions of both cylindrical co-ordinates for the two circular foundations. Then, the variational principle and the reciprocal theorem are employed to generate the dynamic stiffness matrices for the two foundations. In the process of employing the variational principle, a co-ordinate transformation matrix between two cylindrical co-ordinate systems is introduced. Some numerical results of dynamic stiffness matrices for the interaction of two identical rigid circular foundations are presented in order to show the effectiveness and efficiency of the present method, and some elaborations for its future extensions are also discussed.