Room-Temperature-Processable Highly Reliable Resistive Switching Memory with Reconfigurability for Neuromorphic Computing and Ultrasonic Tissue Classification

Reversible metal-filamentary mechanism has been widely investigated to design an analog resistive switching memory (RSM) for neuromorphic hardware-implementation. However, uncontrollable filament-formation, inducing its reliability issues, has been a fundamental challenge. Here, an analog RSM with 3D ion transport channels that can provide unprecedentedly high reliability and robustness is demonstrated. This architecture is realized by a laser-assisted photo-thermochemical process, compatible with the back-end-of-line process and even applicable to a flexible format. These superior characteristics also lead to the proposal of a practical adaptive learning rule for hardware neural networks that can significantly simplify the voltage pulse application methodology even with high computing accuracy. A neural network, which can perform the biological tissue classification task using the ultrasound signals, is designed, and the simulation results confirm that this practical adaptive learning rule is efficient enough to classify these weak and complicated signals with high accuracy (97%). Furthermore, the proposed RSM can work as a diffusive-memristor at the opposite voltage polarity, exhibiting extremely stable threshold switching characteristics. In this mode, several crucial operations in biological nervous systems, such as Ca²⁺ dynamics and nonlinear integrate-and-fire functions of neurons, are successfully emulated. This reconfigurability is also exceedingly beneficial for decreasing the complexity of systems—requiring both drift- and diffusive-memristors.     

The Optimal Spectrum Sensing Time for Maximizing Throughput of 802.11-Based MAC Protocol for Cognitive Radio Networks Under Unsaturated Traffic Conditions

Cognitive radio has attracted considerable attention as an enabling technology for addressing the problem of radio frequency shortages. In cognitive radio networks (CRNs), secondary users (SUs) are allowed to opportunistically utilize the licensed spectrum bands of primary users (PUs) when these bands are temporarily unused. Thus, SUs should monitor the licensed spectrum bands to detect any PU signal. According to the sensing outcomes, SUs should vacate the spectrum bands or may use them. Generally, the spectrum sensing accuracy depends on the sensing time which influences the overall throughput of SUs. That is, there is a fundamental tradeoff between the spectrum sensing time and the achievable throughput of SUs. To determine the optimal sensing time and improve the throughput of SUs, considerable efforts have been expended under the saturated traffic and ideal channel assumptions. However, these assumptions are hardly valid in practical CRNs. In this paper, we provide the framework of an 802.11-based medium access control for CRNs, and we analyze this framework to find the optimal spectrum sensing time under the saturated and unsaturated traffic condition. Through simulation, the proposed analytic model is verified and the fundamental problem of the sensing-throughput tradeoff for CRNs is investigated.     

试样薄化导致的Cu－Al贝氏体／马氏体组织变化

加热到１１７３Ｋ的Ｃｕ－２５ａｔ；％Ａｌ合金块状试样经７２３Ｋ６０Ｓ等温淬火后其组织由片状的“羽毛状贝氏体（９Ｒ结构）和β＿１＇马氏体（１８Ｒ结构）组成。当试样被离子束轰去进一步薄化后，在足够薄的区域内生成了许多２Ｈ结构的薄片马氏体。初步分析认为该薄片马氏体形成的原因是由于贝氏体生成时在试样内留下了高的应力场，当试样薄化到一临界厚度以下时，试样区的三维约束条件松弛使试样内的应力场触发了薄片马氏体的生成。这些薄片马氏体是通过位于贝氏体界面上位错的重新启动而长大的。     

Diversity reception of 16STAR-QAM in frequency-selective Ricianfading

Bit error rate (BER) performance of differential 16STAR-QAM in frequency-selective Rician fading channels with diversity reception is theoretically analysed for three different types of delay profiles: double spike, one-sided exponential and Gaussian profiles. The effect of time delay between line-of-sight (LOS) and multipath components is also included in the analysis, and is shown to degrade the system performance significantly     

An adaptive algorithm for separating and tracking multipledirectional sources in linear arrays

An algorithm for spatially filtering out, enhancing, and tracking individual directional sources in an adaptive array is proposed and investigated. In this algorithm, the sources are separated by using an adaptive beamformer whose outputs are processed by using the LMS algorithm to track distinct sources individually. From the LMS weights used, the source locations can be estimated. Whenever significant changes in these are detected, the beamformer is updated so that its outputs will be due to different sources in the steady state. With this algorithm, the problems of look-direction errors in look-direction constrained arrays and of large signal power in power inversion arrays are eliminated, and the enhancement of multiple moving sources becomes a natural process. Furthermore, because the sources are individually tracked and the beamformer is only updated occasionally, the algorithm possesses fast tracking behavior, and its implementation complexity is comparable to that of beamformer-based adaptive arrays using the LMS algorithm     

Threshold voltage and C-V characteristics of SOIMOSFET's related to Si film thickness variation on SIMOX wafers

C-V characteristics of fully depleted SOI MOSFETs have been studied using a technique for measuring silicon-film thickness using a MOSFET. The technique is based on C-V measurements between the gate and source/drain at two different back-gate voltages, and only a large-area transistor is required. Using this technique, SOI film thickness mapping was made on a finished SIMOX wafer and a thickness variation of ±150 Å was found. This thickness variation causes as much as a 100-mV variation in the device threshold voltage. The silicon-film thickness variation and threshold-voltage variation across a wafer shows a linear correlation dependence for a fully depleted device. C-V measurements of the back-gate device yield the buried-oxide thickness and parasitic capacitances. The effects of GIDL (gate-induced drain leakage) current on C-V characteristics are also discussed     

A new simultaneous noise and input power matching technique formonolithic LNA's using cascode feedback

In this paper, a new simultaneous impedance-matching technique of Γ_opt (optimum noise-match source reflection coefficient) and G_max (maximum available power gain-match (MAPG) source reflection coefficient) using cascode feedback (CF) is proposed. A 1.57-GHz single-stage monolithic-microwave Integrated-circuit (MMIC) low-noise amplifier (LNA) designed with this technique has been fabricated using GaAs MESFET technology in order to verify the feasibility of this scheme. The measured response agrees well with the simulated performance. Extensive computer simulation shows that when silicon npn bipolar junction transistor (BJT) is used, this scheme enables us to make both Γ_opt and G_max points near to 50 Ω, in addition to the simultaneous noise and input power matching. In addition, it has all the advantages of negative feedback such as stability, wider bandwidth, and insensitivity against parameter variation     

A Novel and Efficient Feature Extraction Method for Iris Recognition

With a growing emphasis on human identification, iris recognition has recently received increasing attention. Iris recognition includes eye imaging, iris segmentation, verification, and so on. In this letter, we propose a novel and efficient iris recognition method which employs a cumulative‐sum‐based grey change analysis. Experimental results demonstrate that the proposed method can be used for human identification in efficient manner.     

Solution‐Processable,High‐Performance Flexible Electroluminescent Devices Based on High‐k Nanodielectrics

Flexible alternating‐current electroluminescent (ACEL) devices have attracted considerable attention for their ability to produce uniform light emission under bent conditions and have enormous potential for applications in back lighting panels, decorative lighting in automobiles, and panel displays. Nevertheless, flexible ACEL devices generally require a high operating bias, which precludes their implementation in low power devices. Herein, solution‐processed La‐doped barium titanate (BTO:La) nanocuboids (≈150 nm) are presented as high dielectric constant (high‐k) nanodielectrics, which can enhance the dielectric constant of an ACEL device from 2.6 to 21 (at 1 kHz), enabling the fabrication of high‐performance flexible ACEL devices with a lower operating voltage as well as higher brightness (≈57.54 cd m^?2 at 240 V, 1 kHz) than devices using undoped BTO nanodielectrics (≈14.3 cd m^?2 at 240 V, 1 kHz). Furthermore, a uniform brightness across the whole panel surface of the flexible ACEL devices and excellent device reliability are achieved via the use of uniform networks of crossaligned silver nanowires as highly conductive and flexible electrodes. The results offer experimental validation of high‐brightness flexible ACELs using solution‐processed BTO:La nanodielectrics, which constitutes an important milestone toward the implementation of high‐k nanodielectrics in flexible displays.     

Monolithic integration of microlensed resonant photodetectors andvertical cavity lasers

Monolithic integration of high performance microlensed resonant photodetectors and vertical cavity lasers (VCLs) from a single epitaxial growth is presented. The VCLs have sub-200 μA threshold currents. Adjacent detectors have the same operating wavelength and responsivities of ~0.4 A/W with ~6 nm optical bandwidths