In this paper, the eye-opening conditions are investigated for the minimum bandwidth digital transmission systems where the sinc functions is used as a basic pulse so that only the Nyquist bandwidth is required for the channel. In order for these systems to be tolerant to the timing jitter in the receiver, the eye pattern should be open horizontally. It is proved that an eye opening condition which has been understood only as a sufficiency becomes a necessity as well. As its result, having a spectral null at teh Nyquist frequency is shown to have the same meaning as eye opening.

In this paper, using the principle of duality between the spatial and spectral domain, we proposed a new microwave imaging technique of a coherent tomographic formulation in the spatial domain and reconstructed the image of dielectric cylinder through simulation and experiment. The numerical and experimental results for the variety o object size, relative dielectric constant have shown the limitation of Born approximation to be used and the effect of retrieved images for various signal bandwidth.

In this paper the theoretical analysis of the site attenuation was made on the concept of mutual impedance of antennas and the cause of characteristic hump in the site attenuation curve below 80 MHz is explained. The deviation between the FCC site attenuatios and the experimental results of an ideal open area test site at low frequencies(below 80MHz) was found to be the mutual coupling effects of the antennas under the FCC's measurement conditions. It seems desirable that the site attenuations curve of FCC document be revicused to the theoretical curves presented in this paper is long as the FCC's measurement conditions are to the applied.

A design method of ultra-broadband microwave amplifier is presented. A lossy network is represented as the combination of a serial impedance component and a parallel admittance component, and the realizable ranges of the gain and the reflection coefficients are derived with the components connected to the input, output or interstage network. The matching network has been designed by using the serial and parallel immittance loci which have the constant gain or reflection coefficients within the realizable ranges. Using the proposed method, deisgn examples of ultra-broadband amplifiers operating from dc to 12GHz frequency range are presented.

In this paper, a reliable control system structured with dual CPU modules and dual I/O modules is implemented as a means of achieving a highly reliable fault tolerant control system. For this, faults in the system modules are first examined, and a fault detection technique consisting of self diagnostic, comparison process, and exception processing is applied. Self diagnostic is used to locate which components in the modules have been failed, while comparison process is to cmpare control outputs computed by both CPU modules and protect the plant from malfunction by blocking failed control outputsin advance. Finally exception processing is used to determine the faults that are not detected immediately by the self diagnostic and comparison process, e.g. bus error processing when acknowledge signal for data transfer is not activeted in the I/O modules. Also reliability analysis is conducted for the discrete time Markov model with dual structure. It is shown quantitatively that the reliability is improved in the control system with dual structure in comparison with a system with single module structure.

We propose a time-optimal path planning scheme for the robot manipulator using Hopfield neural network. The time-optimal path planning, which can allow the robot system to perform the demanded tasks with a minimum execution time, may be of consequence to improve the productivity. But most of the methods proposed till now suffers from a significant computational burden and thus limits the on-line application. One way to avoid such a difficulty is to apply the neural networke technique, which can allow the parallel computation, to the minimum time problem. This paper proposes an approach for solving the time-optimal path planning by using Hopfield neural network. The effectiveness of the proposed method is demonstrarted using a PUMA 560 manipulator.

We constructed an intelligent tutoring system that compose four devices i.e. intelligent interface, acqusition module, tutoring controller and tutorial knowledge. We implemented the intelligent tutoring system using prolog as an authoring language in the ultrix os under VAX-11/750, and we propose goal opriented tutoring system algorithm.

In this paper we propose a model which can predict well the logical errors come from the charge redistribution in domino gates. In this model the effect of the intrinsic capacitance between gate and channel of MOSFET's is considered. This effect is more important than the parasitic capacitance effect. The error by the proposed model is only 8% of that by the currently used model. This model can be used as a guide-line in the design of domino circuits.

Characteristics of the direct thermal nitrided films by RF induction heating has been studied. The nitrided films on Si were prepared at 1000-1200\ulcorner in ammonia gas ambient. The nitrided films were analyzed by ellipsometry an Auger electron spectroscopy. I-V and C-V characteristics of MIS capacitors fabricated using nitrided film were investicated. The nitrided films were grown up mostly within initial thirty minutes and no significant growth was observed thereafter. Etch rates of films were about 1\ulcornermin in diluted HF (HF:H2O= 1:50). The nitrided films were resistant to dry and wet oxidations at temperatures below 1000\ulcorner and 900\ulcorner, respectively.

The physical properties of \ulcorner-Fe2O3 raw materials were investigated. \ulcorner-Fe2O3/Pd thick film gas sensors were fabricated with screen-printing method and their electrical and sensitivity characteristics were analyzed. The irreversible phase transition from \ulcorner-Fe2O3 to \ulcorner-Fe2O3 occured at 500\ulcorner. At this time, the cation of tetrahedral sites moved into the octahedral sites. \ulcorner-Fe2O3 raw materials contained only trivalent and no divalent iron. Thecontents of divalent iron (Fe+\ulcorner were increased as detecting gases were adsorbed. The addition of Pd (1w/o) to \ulcorner-Fe2O3 enhanced the sensitivity to gases. The sentivity of \ulcorner-Fe2O3/Pd(1w/o) thick film to 5000ppm C4H10 was 97% at the operating temperature of 300\ulcorner.

We studied the electrical, optical and structural properties of microcrystalline silicon thin films prepared by a new chemical vapour deposition technique, called filament assisted(FA)-CVD. The microcrystalline silicon is sucessfully deposited when the hydrogen dilution ratio exceeds 30. The Raman peak at 520 cm-1 and the X-ray diffraction peak at 27.7\ulcorner0.2\ulcornerbecome sharper with increasing hydrogen dilution ratio. We obtain high quality microcrystalline Si by FA-CVD with optical gap of \ulcorner2.2eV and hydrogen content of \ulcorner3 at %.

Characteristics of NMOS transistors with phosphorus source/drain junctions formed by two-step rapid thermal diffusion (RTD) process using a solid diffusion source have been investigated. Phosphorus profiles after RTD were measured by SIMS analysis. In the case of 1100\ulcorner, 10sec RTD of, P, the specific contact resistance of n+ Si-Al was 2.4x10**-7 \ulcorner-cm\ulcorner which is 1/5 of the As junction The comparison fo P junction devices formed by RTD and conventional As junction devices shows that both short channel effect and hot carrier effect of P junction devices are smaller than those of As junction devices when the devices have same junction depths. P junction device had maximum of 0.4 times lower Isub/Id than As junction device. Characteristics of P junction formed by several different RTD conditions have been compared and 1000\ulcorner RTD sample had the smaller hot carrier generation. Also, it has been shown that the hot carrier generation can be futher reduced by forming the P junctions by 3-step RTD which has RTO-driven-in process additionally.

The microstructure and electrical conductivity characteristics were examined for the Yttria Stablized Zirconia that were added Titania with the concentration of 0.02wt%, 0.05wt% and 0.1wt%. The stabilized cubic phase had not been changed. The obtained density of the sintered body was higher than 97% of the theoretical density as the grain growth and proe removal. And the highest density of 5.91g/cm**3 was obtained for 0.02wt% addition of TiO2. When 0.02 wt% TiO2 was added to YSZ, the electrical conductivity was 0.293.(\ulcorner.cm)**-1 and the limit of electrolyte regime Pn was 4x10**-28 atm. at 1, 100\ulcorner, respectively. Those were improved value compared with the electrical conductivity 0.107 (\ulcorner.cm)**-1 and Pn 1.5x10**-27atm. without adding TiO2.

This paper proposes the Selected Harmonic Elimination Pluse Width Modulation (SHE PW) scheme toreduce the ahrmonic components of source line current. To eliminate the low order harmonics which affects the source dominatly, we apply the Fourier series analysis to line current waveforms and then find out the switching patterns using the SHE PWM scheme. In addition to the analysis of harmonic effects, the three phase filter circuit is used to reduce high order harmonics. For the experimental realization, the converter circuit with power Transistor(PTR) is designed and the Pulse Time Control(PTC) is applied. The line current and the load voltage are measured under the condition of three phase application, highly inductive load.

Prototype Switched Capacitor 4th order Low pass and Band Pass Filters were Realized and tested. Their capacitor values were determined by using automatic SCF design tool. Each filter was designed to operate in voiceband frequency, and practical solutions to improve its frequency characteristics were given. Experimental results show the validity of automatic SCF design program. Simulation and experimental results were fully compared and optimum design conditions were summarized.

In this paper we propose the search algorithm of the boundary nodes that defined as the circuit segmentation when CUT is tested by pseudo-exhaustive testing. The algorithm treats the testability values of the nodes in CUT as the population composed of teh raw data, and after examines the level of significance a and then estimate the confidence interval of teh testability values. Thus One can easily searched the c9oundary nodes and PO of sub circuits. The proposed algorithm has been implemented under UNIX OS with C-language, applied to the combinational logic CUT. As a result, it is shown that the pseudo-exhaustive test patterns are least generated when \ulcornerhas 0.786. We confirmed that the rate of test pattern is 1.22%, compared with exhaustive test.

In this paper, a range image segmentation method is proposed. This method consists of an initial segmentation stage by discontinuous edge detection and surface type labeling based on the sign of the principal curvatures. Initially type labeled image is oversegmented, this image is merged via stepwise optimal region merging stage based on polynomial function approxiamtion. The successful segmentation results are presented for two synthetic range images with noise and a real-world ERIM range image.

The time domain adaptive noise canceller (time domain ANC) with the adaptive compensator and its algorithm, so called compensated least mean squares(CLMS) algorithm, had been introduced to improve the performance of ANC[1]. In this paper the time domain ANC with the adaptive compensator is transformed into the frequency domain ANC with the adaptive ocmpensator. An compensated frequency-domain least mean squares(CFLMS) algorithm that can adapt the proposed frequency domain ANC is presented.

In this paper, isolated words are recognized using multisection VQ and HMM. As recognition vocabuaries, 20 area-name which is uttered 5 times by 3 speakers is selected. In generating codebook, we devide recognition vocabulary into equal length, section, and make standard VQ codebook to each section and calculate observation by section and than recognize isolated words by HMM training. Multisection VQ codebook has time information and as observation is calculated by eacy section, computation is lesser and recongnition rate is higher than by whole codword. As a result, it is proved that recognition rate is higher in case of HMM using multisection VQ codebook.