For a multi-rate serial link, the proposed TX rise/fall time (tTR) control scheme enables tTR to be changed according to data rate. The main driver is split into multiple legs and each leg is controlled by its own pre-driver. All pre-drivers are biased on for the smallest tTR while some pre-drivers are biased off for large tTR. Due to RC delay among pre-drivers, the transition of a biased-off pre-driver's output occurs later but becomes overlapped with those of adjacent pre-drivers' outputs so that there is no stair-step waveform shown. The measurement results show that the driver has a sufficient tTR control range of around three times.

抄録全体を表示

A 10 Gbps transmitter bridge chip including four data lanes, which increases the bandwidth using an 8-to-1 serialization, is proposed for a field-programmable gate array (FPGA)-based frame generator to support the protocol of the D-PHY version 1.2 for the mobile industry processor interface (MIPI) display serial interface (DSI).

抄録全体を表示

A 3-Gbps decision feedback equalizer (DFE) compensating for data and edge inter-symbol interference (ISI) is presented. A speculative architecture is employed to relieve the timing burden on the feedback signal for the DFE wherein the ISI of edge sample is compensated by speculating the DFE based on two-UI earlier data sample. Thereby, the timing margins of the DFE for data and edge ISI compensation are ensured to be larger than 1.0-UI. The proposed DFE has been implemented in a 0.13-µm CMOS technology together with a clock and data recovery (CDR) circuit. The DFE and CDR circuits occupy 0.28-mm² active area and the DFE consumes 18-mW from a 1.2-V supply. The RMS jitter of the recovered clock is improved from 15.6-ps to 11.9-ps by the proposed edge ISI compensating DFE.

抄録全体を表示

This paper presents the results from an experimental investigation in the near field characteristics and the flow interaction of a doublet flow which consists of coaxial injection and suction flow. The mixing and diffusion of the inner injection flow can be manipulated by the outer suction flow. The higher the ratio of the suction velocity U _s at the pipe edge to the injection one U _i at the same location, the faster the decay of the center axis mean velocity in the doublet flow becomes. For a large value of U _s/U _i, the fluctuation velocity on the center axis abruptly increases and the doublet flow behaves an oscillating turbulent flow. From the results of flow visualization, it is clear that the width and the spreading angle of the doublet flow increase rapidly with increasing U _s/U _i and are larger than that of a free jet.

抄録全体を表示

This article presents a study of incorporation and the use of incorporated words in the interactions of nonnative (Japanese) learners at the lower intermediate and the upper-intermediate oral proficiency levels. Transcripts totaling about 160 minutes of pair's interaction taken from a shape-description task were analyzed using some indexes developed for this study according research questions. Results reveal that the subjects at the lower-intermediate level (LI) were less likely to incorporate words from utterances by the interlocutor in contrast to the subjects at the upper-intermediate level (UI). Furthermore, there were overwhelming tendencies for LIs to modify incorporated words with no requirement, not to modify them with requirement, and to repeat them without attaching them to their original utterances during interaction. Careful consideration of the use of incorporated words in later interactions shed light on the tendencies that they were hardly used in all of the interactions and were easily replaced with words which were in input and had close meaning in ongoing interaction. The findings indicate the lack of durability of the incorporated words and the disadvantage for second language acquisition (SLA).

抄録全体を表示

In this study, the direct numerical simulation (DNS) for homogeneous shear turbulence in the system rotating along the streamwise direction is fulfilled. Due to the rotation effect, the turbulence energy becomes small during the initial short time and the Reynolds shear stress are suppressed more strongly with increasing the system angular velocity. Since the redistribution related to the rotation and pressure-strain terms of the Reynolds normal stress is weakened by the streamwise rotation, the anisotropy of the normal stresses is strengthened. In the strong rotation case the anisotropy of the velocity spectra is strong in the whole wavenumber region in contrast with the isotropy of the nonrotating case in the large one. From the viewpoints of the probability density function (PDF) for the vorticity vector angle and visualization for the vortex structure, we find that the vortex structures become large and stand in a line by the streamwise rotation axis. Moreover, we suggest that the rapid distortion theory (RDT) simulation reproduces the rotation effect on the mean quantities of the DNS results in only short time immediately after a calculation start.

抄録全体を表示

This paper presents a 36 Gb/s receiver equalizer including an adaptive continuous time linear equalizer (CTLE), which is based on slope detection and a half-rate speculative decision feedback equalizer (DFE) in 0.13 µm BiCMOS technology for high speed serial link. The CTLE with middle frequency compensation can not only adjust the ratio of high frequency and low frequency components adaptively, but also provide a small amount of equalization to middle frequency range. A half-rate speculative DFE, which is connected to the back of the CTLE, can satisfy the time constraints and eliminate the residual inter-symbol interference. The chip area including pads is about 1.2 mm² and the power consumption is about 750 mW under 3.3 V power supply. Measurement results show that the receiver chip can effectively equalize 24 dB loss at Nyquist frequency and a clear eye diagram can be captured at 36 Gb/s.

抄録全体を表示

This paper presents a channel operating margin (COM) based high-speed serial link optimization using machine learning (ML). COM that is proposed for evaluating serial link is calculated at first and during the calculation several important equalization parameters corresponding to the best configuration are extracted which can be used for the ML modeling of serial link. Then a deep neural network containing hidden layers are investigated to model a whole serial link equalization including transmitter feed forward equalizer (FFE), receiver continuous time linear equalizer (CTLE) and decision feedback equalizer (DFE). By training, validating and testing a lot of samples that meet the COM specification of 400GAUI-8 C2C, an effective ML model is generated and the maximum relative error is only 0.1 compared with computation results. At last 3 link configurations are discussed from the view of tradeoff between the link performance and cost, illustrating that our COM based ML modeling method can be applied to advanced serial link design for NRZ, PAM4 or even other higher level pulse amplitude modulation signal.

抄録全体を表示

In the slab reheating process, the temperature distribution of walking beam furnace is critical for the temperature of slab, as well as the subsequent hot rolling. Besides, this process is dynamic, nonlinear, and time varying due to coupling with complex physical and chemical reactions. In this paper, we focus on an operation optimization problem of tuning furnace temperature to optimize the slab reheating process, which considers the reheating quality and production cost. For this purpose, we develop a novel operation optimization method. Specifically, first, from the view of mechanism, a heat transmit model based on the heat-exchange of slab and furnace, an energy consumption function, and especially, an oxidation loss function are established. Therefore, incorporating the mechanism models, we build an optimization model to minimize temperature deviation of slab, energy consumption, and oxidation loss to describe our problem. Then, based on the features of the problem, we propose a modified differential evolution algorithm, which includes a space contraction scheme, a new self-adaptive parameter strategy, and a new mutation operator. Finally, to evaluate the performance of our method, extensive numerical experiments are implemented by comparing it with other well-known evolutionary algorithms based on practical data. The experimental results demonstrate the effectiveness of proposed operation optimization method on solving our problem.

抄録全体を表示

The phonon contribution to the resistive force acting on a moving dislocation is discussed, and an expression of the dislocation damping constant is obtained. The damping constant of a dislocation in an isotropic medium is expressed analytically. The theoretical results are also applied to anisotropic alkali halide crystals, and the numerical results are compared with the experimental results.

抄録全体を表示

A large number of embedded computers, such as network appliances and sensors, have rapidly spread out to home and office environments in the last few years. These embedded computers have enough CPU power to execute the software components that can control hardware. Managing distributed components together can enhance human activity and change the real world into a “Smart Space.” We name such collaboration of components “federated service” or “application.” In this paper, we have developed and evaluated a novel middleware named uBlocks which enables users to build and manage applications. uBlocks, unlike other distributed application-building middleware, is distinguished by two major features. The first is a flexible communication mechanism named RT/Dragon. RT/Dragon enables the connection of heterogeneous components. The second is the universal modeling of various distributed components to support building applications by multiple users in parallel. Additionally, to enable building applications in a simple way, we provide various user interfaces (UI) for multi-modal visualization: 2D/3D User Interface, and a web interface. These features lead to reduce the cost of building and managing distributed applications by the user. This research proves that the idea of building applications by users is practical and effective.

抄録全体を表示

A large number of embedded computers, such as network appliances and sensors, have rapidly spread out to home and office environments in the last few years. These embedded computers have enough CPU power to execute the software components that can control hardware. Managing distributed components together can enhance human activity and change the real world into a “Smart Space.” We name such collaboration of components “federated service” or “application.” In this paper, we have developed and evaluated a novel middleware named uBlocks which enables users to build and manage applications. uBlocks, unlike other distributed application-building middleware, is distinguished by two major features. The first is a flexible communication mechanism named RT/Dragon. RT/Dragon enables the connection of heterogeneous components. The second is the universal modeling of various distributed components to support building applications by multiple users in parallel. Additionally, to enable building applications in a simple way, we provide various user interfaces (UI) for multi-modal visualization: 2D/3D User Interface, and a web interface. These features lead to reduce the cost of building and managing distributed applications by the user. This research proves that the idea of building applications by users is practical and effective.

抄録全体を表示

Preference dependency grammar (PDG) is a framework for integrating morphological, syntactic, and semantic analyses. PDG provides packed shared data structures that can efficiently encompass all possible interpretations at each level of sentence analyses with preference scores. Using the structure, PDG can calculate a globally optimized interpretation for the target sentence. This paper first gives an overview of the PDG framework by describing the base model of PDG, which is a sentence analysis model, called a “multi-level packed shared data connection model.” Then this paper describes packed shared data structures, e.g., headed parse forests and dependency forests, adopted in PDG. Finally, the completeness and soundness of the mapping between the parse forest and the dependency forest are revealed.

抄録全体を表示