Voyager Neptune radio science data were collected using three antennas on Earth on August 25, 1989. A parabolic antenna at Canberra, Australia, of 70 meter diameter received 2.3GHz and 8.4GHz carriers. The 64 meter parabolic antennas at Parkes. Australia and Usuda, Japan, received only the 8.4GHz and only the 2.3GHz carriers, respectively. It is necessary to reduce the frequency variation in the received signal carrier to extract accurate information on physically interesting objects such as Neptune's atmosphere, ionosphere, or the rings. After the frequency stabilization process, the frequency drift was reduced from 180Hz down to a maximum of 5Hz, making it possible to reduce the data bandwidth and, consequently, the data volume, by a factor of 30. The uncertainty of the signal frequency estimates were also reduced from 5 down to 510-3Hz/sec above the atmosphere, from 5 down to 0.5Hz/sec in the atmosphere, and from 50 down to 3Hz/sec at the beginning and the end of the atmospheric occultation. Much of the remaining uncertainty is due to scintillations in Neptune's atmosphere and cannot be reduced further. The estimates are thus meaningfully accurate and suitable for scientific analysis and coherent arraying of data from different antennas. Two results based on these estimates are shown: a preliminary temperature-pressure (T-p) profile of Neptune's atmosphere down to a pressure level of 2 bar computed using the Usuda 2.3GHz data, and a multipath phenomenon in the atmosphere seen in Canberra 8.4GHz data. Our T-p profile shows good agreement with the results presented by Lindal et al. within 1K below 100mbar pressure level, even though our result is based on an independent data set and processing. A comparison of the multipath phenomena at Neptune with that at Uranus implies that it was created by a cloud layer with a smaller scale height than the atmosphere above and below it. The processing methods described were developed in part with the interest to coherently array Canberra, Parkes and Usuda data. In this sense, while this paper does not extend any science results, the observations and results are derived independently from other published results, and in the case of Usuda, are completely new.

The mechanism of UV-excited dry cleaning using photoexcited chlorine radicals has been investigated for removing iron and aluminum contamination on a silicon surface. The iron and aluminum contaminants with a surface concentration of 1013 atoms/cm2 were intentionally introduced via an ammonium-hydrogenperoxide solution. The silicon etching rates from the Uv-excited dry cleaning differ depending on the contaminants. Fe and Al can be removed in the same manner. The removal of Fe and Al is highly temperature dependent, and is little affected by the silicon etching depth. Both Fe and Al on the silicon surface were completely removed by UV-excited dry cleaning at a cleaning temperature of 170, and were decreased by two orders of magnitude from the initial level when the surface was etched only 2 nm deep.

This paper designs and evaluates highly parallel VLSI processors for real time 2-D state-space digital filters using hierarchical behavioral description language and synthesizer. The architecture of the 2-D state-space digital filtering system is a linear systolic array of homogeneous VLSI processors, each of which consists of eight processing elements (PEs) executing 1-D state-space digital filtering with multi-input and multi-output. Hierarchical behavioral description language and synthesizer are adopted to design and evaluate PE's and the VLSI processors. One 16 bit fixed-point PE executing a (4, 4)-th order 2-D state-space digital filtering is described on the basis of distributed arithmetic in about 1,200 steps by the description language and is composed of 15 K gates in terms of 2 input NAND gate. One VLSI processor which is a cascade connection of eight PEs is composed of 129 K gates and can be integrated into one 1515 [mm2] VLSI chip using 1 µm CMOS standard cell. The 2-D state-space digital filtering system composed of 128 VLSI processors at 25 MHz clock can execute a 1,0241,024 image in 1.47 [msec] and thus can be applied to real-time conventional video signal processing.

It is crucial to make Si wafer surfaces ultraclean in order to realize such advanced processes as the low-temperature process and the high-selectivity in the ULSI production. The ultra clean wafer surface must be perfectly free from particles, organic materials, metallic impurities, native oxide, surface microroughness, and adsorbed molecule impurities. Since the metallic contamination on the wafer surface, which is one of the major contaminants to be overcome in order to come up with the ultra clean wafer surface, has the fatal effect on the device characteristics, the metallic impurities in the wafer surface must be suppressed at least below 1010 atoms/cm2. Meanwhile the current dry processes such as reactive ion etching or ion implantation, suffer the metallic contamination of 10121013 atoms/cm2. The wet process becomes increasingly important to remove the metallic impurities introduced in the dry process. Employing a new evaluation method, the metallic impurity segregations at the inrerface between the Si and liquid employed in the wet cleaning process of the Si surface such as ultrapure water and various clemicals were studied. This article clearly indicate that it is important to suppress the metallic impurities, such as Cu, which can exchange electrons with Si to be segregated, at least below the 10 ppt level in ultrapure water and liquid chemical such as HF, H2O2, which are employed in the final step of the wet cleaning. When the ultrapure water rinsing is performed in the ambience containing oxygen, the native oxide grows accompanying an inclusion of metals featuring lower electron negativity than Si. It is revealed that, in order to provent the metallic impurity precipitation, it is require not only to remove metallic impurities from ultrapure water but also to keep the cleaning ambience without oxygen, such as the nitrogen ambience, so as to suppress the native oxide formation.

This letter discusses the quality improvement of reconstructed images in diffraction tomography. An efficient iterative procedure based on the modified Newton-Kantorovich method and the Gerchberg-Papoulis algorithm is presented. The simulated results demonstrate the property of high-quality reconstruction even for cases where the first-order Born approximation fails.

This paper presents a system that automatically refines the theory expressed in the function-free first-order logic. Our system can efficiently correct multiple faults in both the concept and subconcepts of the theory, given only the classified examples of the concept. It can refine larger classes of theory than existing systems can since it has overcome many of their limitations. Our system is based on a new combination of an inductive and an explanation-based learning algorithms, which we call the biggest-first multiple-example EBL (BM-EBL). From a learning perspective, our system is an improvement over the FOIL learning system in that our system can accept a theory as well as examples. An experiment shows that when our system is given a theory that has the classification error rate as high as 50%, it can still learn faster and with more accuracy than when it is not given any theory.

Logarithmic number systems (LNS) provide a very fast computational method. Their exceptional speed has been demonstrated in signal processing and then in computer graphics. But the precision problem of LNS in computer graphics has not been fully examined. In this paper analysis is made for the problem of LNS in picture generation, in particular for circle drawing. Theoretical error analysis is made for the circle drawing. That is, some expressions are developed for the relative error variances. Then they are examined by simulation experiments. Some comparisons are also done with floating point arithmetic with equivalent word length and dynamic range. The results show that the theory and the experiments agree reasonably well and that the logarithmic arithmetic is superior to or at least comparable to the corresponding floating point arithmetic with equivalent word length and dynamic range. Those results are also verified by visual inspections of actually drawn circles. It also shows that the conversion error (from integer to LNS), which is inherent in computer graphics with LNS, does not make too much influence on the total computational error for circle drawing. But it shows that the square-rooting makes the larger influence.

The meaning of analogical reasoning in locally stratified logic programs are described by generalized stable model (GSM) semantics. Although studies on the theoretical aspects of analogical reasoning have recently been on the increase, there have been few attempts to give declarative semantics for analogical reasoning. This paper takes notice of the fact that GSM semantics gives meaning to the effect that the negated predicates represent exceptional cases. We define predicates that denote unusual cases regarding analogical reasoning; for example, ab(x)p(x)g(x), where p(s), q(s), p(t) are given. We also add rules with negated occurrences of such predicates into the original program. In this way, analogical models for original programs are given in the form of GSMs of extended programs. A proof procedure for this semantics is presented. The main objective of this paper is not to construct a practical analogical reasoning system, but rather to present a framework for analyzing characteristics of analogical reasoning.

Effects of cleaning by H2SO4: H2O2 on thin SiO2 film was investigated. The cleaning increases Fowler-Nordheim currents by about 14%, shifts the dielectric breakdown distribution to lower electric field intensity and degrades TDDB characteristics. These results are due to the oxidation and commensurate roughening of the silicon srface by the cleaning solution. When the cleaning is done at higher temperature and with higher concentration of hydroperoxide, microroughness of silicon surface increases. Therfore, the trade-off between the cleaning effect and the roughening effect of H2SO4: H2O2 should be found out.

Based on the Fornasini-Marchesini second model, an efficient algorithm is developed to derive the characteristic polynomial and the inverse of the system matrix from the state-space parameters. As a result, the external description of the Fornasini-Marchesini second model is clarified. A technique for designing 2-D recursive digital filters in the frequency domain is then presented by using the Fornasini-Marchesini second model. The resulting filter approximates both magnitude and group delay specifications and its stability is always guaranteed. Finally, three design examples are given to illustrate the utility of the proposed technique.

We present a successful method for designing 2-D circularly symmetric R lowpass filters with constant group delay. The procedure is based on a transformation of a 1-D prototype R filter with constant group delay, whose magnitude response is the 2-D cross-sectional response. The 2-D filter transfer function has a separable denominator and a numerator which is obtained from the prototype numerator by means of a series of McClellan transformations whose free parameters can be optimized by successive procedure. The method is illustrated by an example.

We introduce a novel neural network called the T-Model and investigates the learning ability of the T-Model neural network. A learning algorithm based on the least mean square (LMS) algorithm is used to train the T-Model and produces a very good result for the T-Model network. We present simulation results on several practical problems to illustrate the efficiency of the learning techniques. As a result, the T-Model network learns successfully, but the Hopfield model fails to and the T-Model learns much more effectively and more quickly than a multi-layer network.

In this paper, subband image coding with symmetric biorthogonal wavelet filters is studied. In order to implement the symmetric biorthogonal wavelet basis, we use the Laplacian Pyramid Model (LPM) and the trigonometric polynomial solution method. These symmetric biorthogonal wavelet basis are used to form filters in each subband. Also coefficients of the filter are optimized with respect to the coding efficiency. From this optimization, we show that the values of a in the LPM generating kernel have the best coding efficiency in the range of 0.7 to 0.75. We also present an optimal bit allocation method based on considerations of the reconstruction filter characteristics. The step size of each subband uniform quantizer is determined by using this bit allocation method. The coding efficiency of the symmetric biorthogonal wavelet filter is compared with those of other filters: QMF, SSKF and Orthonormal wavelet filter. Simulation results demonstrate that the symmetric biorthogonal wavelet filter is useful as a basic means for image analysis/synthesis filters and can give better coding efficiency than other filters.

The two most important parameters in reactive ion etching process, ion bombardment energy and flux, were extracted through a simple RF waveform measurement at the excitation electrode in a conventional cathode-coupled plasma RIE system. By using the extracted plasma parameters, damage and contamination in Si substrates induced by reactive ion etching in a SiCl4 plasma were investigated. A very convenient map representation of ion energy and ion flux was introduced in understanding the etching process occurring in the RIE system.

This paper describes a method of evaluating operations effort for fiber optic subscriber loops, such as the Central Terminal/Remote Terminal (CT/RT) system, which can economically provide a variety of telecommunication services. Four system configurations with different operation procedures are evaluated by simulation. By evaluating the operating costs associated with service provisioning, it is shown that automatic distributing frames are cost effective in subscriber loops with CT/RT systems. Moreover, the most economical operation strategies for installing and extending subscriber boards are discussed in terms of facility and operations cost.

In this paper, restricted overflow strategy is proposed as a novel channel access strategy for the queueable hierarchical channel structure, which has been proposed as one of "Wideband-ISDN" channel structures. In this policy, overflow from higher bit rate channels to lower bit rate channels is partly restricted by the number of waiting customers in the higher channel's buffer. Therefore, thresholds, which restrict overflow, are considered on the buffer. First, we present the system model with two types of services and restricted overflow strategy. Next, we provide a queueing analysis of this strategy. After that, some numerical results of both conventional overflow strategy and restricted overflow strategy are presented, and we compare the average holding times under these strategies. Finally, we show that, if we choose appropriate thresholds, the average holding time of higher level traffic is improved.

The increase of surface microroughness on Si substrate degrades the electrical characteristics such as the dielectric breakdown field intensity (EBD) and charge to break-down (QBD) of thin oxide film. It has been found that the surface microroughness increases in the wet chemical process, particularly in NH4OH-H2O2-H2O cleaning (APM cleaning). It has been revealed that the surface microroughness does not increase at all if the NH4OH mixing ratio in NH4OH-H2O2-H2O solution is reduced from the conventional level of 1:1:5 to 0.05:1:5, and the room temperature ultrapure water rinsing is introduced right after the APM cleaning. At the same time, the APM cleaning with NH4OH-H2O2-H2O mixing ratio of 0.05:1:5 has been very effective to remove particles and metallic impurities from the Si surface. The surface microroughness dominating the electrical properties of very thin oxide films is strictly influenced by the wafer quality. The increase of surface microroughness due to the APM cleaning has varied among the wafer types such as Cz, FZ and epitaxial (EPI) wafers. The increase of surface microroughness in EPI wafer was very much limited, while the surface microroughness of FZ and Cz wafers gradually increase. As a result of investigating the amount of diffused phosphorus atoms into these wafers, the increase of the surface microroughness in APM cleaning has been confirmed to strongly depend on the silicon vacancy cluster concentration in wafer. The EPI wafer having low silicon vacancy concentration is essentially revealed superior for future sub-half-micron ULSI devices.

The effects of changing system parameters on job scheduling policies are studied for load balancing of multi-class jobs in a distributed computer system that consists of heterogeneous host computers connected by a single-channel communications network. A job scheduling policy decides which host should process the arriving jobs. We consider two job scheduling policies. The one is the overall optimal policy whereby jobs are scheduled so as to minimize the overall mean job response time. Tantawi and Towsley obtained the algorithm that gives the solution of the policy in the single class job environment and Kim and Kameda extended it to the multiple job class environment. The other is the individually optimal policy whereby jobs are scheduled so that every job may feel that its own expected response time is minimized. We can consider three important system parameters in a distributed computer system: the communication time of the network, the processing capacity of each node, and the job arrival rate of each node. We examine the effects of these three parameters on the two load balancing policies by numerical experiment.

We investigate the relationship between two different notions of reducibility among prediction (learning) problems within the distribution-free learning model of Valiant (PAC learning model). The notions of reducibility we consider are the analogues for prediction problems of the many-one reducibility and of the Turing reducibility. The former is the notion of prediction preserving reducibility developed by Pitt and Warmuth, and its generalization. Concerning these two notions of reducibility, we show that there exist a pair of prediction problems A and B, whose membership problems are polynomial time solvable, such that A is reducible to B with respect to the Turing reducibility, but not with respect to the prediction preserving reducibility. We show this result by making use of the notion of a class of polynomially sparse variants of a concept representation class. We first show that any class A of polynomially sparse variants of another class B is reducible to B with respect to the Turing reducibility'. We then prove the existence of a prediction problem R and a class R of polynomially sparse variants of R, such that R does not reduce to R with respect to the prediction preserving reducibility.

In the approximate learning model introduced by Valiant, it has been shown by Blumer et al. that an Occam algorithm is immediately a PAC-learning algorithm. An Occam algorithm is a polynomial time algorithm that produces, for any sequence of examples, a simple hypothesis consistent with the examples. So an Occam algorithm is thought of as a procedure that compresses information in the examples. Weakening the compressing ability of Occam algorithms, a notion of weak Occam algorithms is introduced and the relationship between weak Occam algorithms and PAC-learning algorithms is investigated. It is shown that although a weak Occam algorithm is immediately a (probably) consistent PAC-learning algorithm, the converse does not hold. On the other hand, we show how to construct a weak Occam algorithm from a PAC-learning algorithm under some natural conditions. This result implies the equivalence between the existence of a weak Occam algorithm and that of a PAC-learning algorithm. Since the weak Occam algorithms constructed from PAC-learning algorithms are deterministic, our result improves a result of Board and Pitt's that the existence of a PAC-learning algorithm is equivalent to that of a randomized Occam algorithm.