共源-共栅组态S2I电流存储单元及其性能

针对原型S2I开关电流存储单元性能上的一些弱点,提出了共源-共栅组态的S2I电流存储单元(简称CS2I)新结构,使其关键速度与精度性能得到较好的改善.相同器件尺寸下的S2I与CS2I单元电路相比,后者速度性能提高了1.6倍,两种电路结构同样应用于延迟单元和双采样双线性积分器功能部件的HSPICE仿真表明:CS2I方式组成的延迟单元的精度提高了5倍,双采样双线性积分器的三次谐波减少了15dB.     

Active control schemes for aseismic base-isolated structures

Summary  Active control schemes are used for the protection of base-isolated and seismically excited buildings. The desired control objective is to keep the whole structure arbitrarily close to its initial configuration prior to the earthquake. The proposed methods require control force application only at the base of the structure. The controllers developed may depend on information obtained from all the floors or just the first (base) floor alone. Received 13 September 2000; accepted for publication 26 June 2001 RID=" ID=" Dedicated to the memory of Professor P.D. Panagiotopoulos with our warmest prayers. RID=" ID=" The first author wishes to thank Prof. G. Leitmann and Prof. E. Papadopoulos for the numerous helpful discussions. The same author is supported in part by the Institute of Communication and Computer Systems, NTUA, under the program Archimedes 65/1017.     

基于3D ResNet-LSTM的多视角人体动作识别方法

在基于视频图像的动作识别中,由于固定视角相机所获取的不同动作视频存在视角差异,会造成识别准确率降低等问题。使用多视角视频图像是提高识别准确率的方法之一,提出基于三维残差网络（3D Residual Network,3D ResNet）和长短时记忆（Long Short-term Memory,LSTM）网络的多视角人体动作识别算法,通过3D ResNet学习各视角动作序列的融合时空特征,利用多层LSTM网络继续学习视频流中的长期活动序列表示并深度挖掘视频帧序列之间的时序信息。在NTU RGB+D 120数据集上的实验结果表明,该模型对多视角视频序列动作识别的准确率可达83.2%。     

基于关系图卷积神经网络的多标签事件预测

事件预测需要综合考虑的要素众多,现有预测模型多数存在数据稀疏、事件的组合特征及时序特征考虑不足、预测类型单一等问题。为此,提出了基于关系图卷积神经网络的多标签事件预测方法,通过节点特征聚合技术实现数据的稠密化表示。模型利用卷积神经网络的卷积和池化运算,提取预测数据的组合时间段特征信息,并结合长短期记忆网络的时序特征提取能力,进一步提取预测数据的时序规律特征;最后,模型通过全连接的多标签分类器,输出多种类型事件发生的概率值。实验结果表明,所提模型可以支持进行多日期、多类型事件预测,在特定数据集上最高F1值可以达到0.85。     

Thermal-Triggered “On–Off” Switchable Triboelectric Nanogenerator Based on Two-Way Shape Memory Polymer

Developing multifunctional triboelectric nanogenerators (TENGs) with special intelligence is of great significance for next-generation self-powered electronic devices. However, the relevant work on the intelligent TENGs, especially those spontaneously responsive to external stimuli, is rarely reported. Herein, an intelligent TENG with thermal-triggered switchable functionality and high triboelectric outputs is developed by designing a movable triboelectric layer, which is driven by a two-way shape memory polyurethane. The resultant TENG device can be spontaneously switched on/off in response to the environmental temperature change, i.e., switching on at 0 °C and off at 60 °C. At the “on” state, the developed TENG exhibits excellent triboelectric performance with a maximum output power density of 5.15 W m⁻² at a pressure of 30 kPa due to the unique advantages of micro-/nanofiber triboelectric surfaces. Furthermore, the great potential of the switchable TENG in intelligent wearable electronic applications is demonstrated, which can serve as not only the sensing element for monitoring human movement and physical condition in a cold environment but also the thermal-driven switch for turning on/off the heating function on demand. The intelligent “on–off” switchable TENG combined with excellent triboelectric performance may provide new opportunities for future self-powered wearable electronics.     

Bio-Inspired Semi-Active Safeguarding Design with Enhanced Impact Resistance via Shape Memory Effect

Inspired by creatures reducing impact damage, a novel semi-active protection strategy to enhance impact resistance through structural adjustment is proposed. The heat-activated shape memory composites (PCLE) consisting of polycaprolactone (PCL) and shear stiffening elastomer are first prepared. The PCLE with 50 wt.% PCL (PCLE-50%) presents tensile strength of 0.46 MPa, fracture strain of 148%, high shape fixity and recovery rates (97.9% and 91.1%), and stable shape-memory repeatability. Furthermore, various PCLE-50%-based structures are fabricated, which show reinforced anti-impact performance after shape-shifting. Loading from 300 mm, the maximum drop hammer impact force of no protection, PCLE-50%, and deformed assembled single-branch structure are 3.98 kN, 1.75 kN, and 0.93 kN, respectively. Additionally, assembled complex structures also exhibit exceptional energy absorption properties. Ultimately, intelligent clothing is successfully manufactured. Besides good anti-impact effect, the clothing shows excellent thermal insulation capabilities whose surface temperature is 31.7 °C after heating for 600 s at 42 °C. This study provides an innovative means to develop deformable, wearable, and intelligent protective devices.     

Bioinspired Artificial Visual System Based on 2D WSe2 Synapse Array (Adv. Funct. Mater. 41/2023)

Machine vision systems that capture images for visual inspection and recognition tasks must be able to perceive, memorize, and compute any color scene. To achieve this, most of the current visual systems use circuits and algorithms which may reduce efficiency and increase complexity. Herein, a 2D semiconductor tungsten diselenide (WSe₂)-based phototransistor that successfully demonstrates an artificial vision system integrating the processing capability of visual information sensing memory, is reported. Furthermore, based on a 6 × 6 fabricated retinal perception array, artificial visual information sensing memory and processing system are proposed to perform image recognition tasks, which can avoid the time delay and energy consumption caused by data conversion and movement. On the other hand, highly linear symmetric synaptic plasticity can be achieved based on the modulation of carrier types in WSe₂ transistors with different thicknesses, facilitating the high level of training and inference accuracy for artificial neural networks. Last, through training and inference simulations, the feasibility of the hybrid synapses for optical neural networks (ONN) is demonstrated.     

Analysis of read speed latency in 6T-SRAM cell using multi-layered graphene nanoribbon and cu based nano-interconnects for high performance memory circuit design

In this study, we designed a 6T-SRAM cell using 16-nm CMOS process and analyzed the performance in terms of read-speed latency. The temperature-dependent Cu and multilayered graphene nanoribbon (MLGNR)-based nano-interconnect materials is used throughout the circuit (primarily bit/bit-bars [red lines] and word lines [write lines]). Here, the read speed analysis is performed with four different chip operating temperatures (150K, 250K, 350K, and 450K) using both Cu and graphene nanoribbon (GNR) nano-interconnects with different interconnect lengths (from 10 μm to 100 μm), for reading-0 and reading-1 operations. To execute the reading operation, the CMOS technology, that is, the16-nm PTM-HPC model, and the16-nm interconnect technology, that is, ITRS-13, are used in this application. The complete design is simulated using TSPICE simulation tools (by Mentor Graphics). The read speed latency increases rapidly as interconnect length increases for both Cu and GNR interconnects. However, the Cu interconnect has three to six times more latency than the GNR. In addition, we observe that the reading speed latency for the GNR interconnect is ~10.29 ns for wide temperature variations (150K to 450K), whereas the reading speed latency for the Cu interconnect varies between ~32 ns and 65 ns for the same temperature ranges. The above analysis is useful for the design of next generation, high-speed memories using different nano-interconnect materials.     

Two-Photon Polymerized Shape Memory Microfibers: A New Mechanical Characterization Method in Liquid

Two-photon polymerization (TPP) is widely used to create 3D micro- and nanoscale scaffolds for biological and mechanobiological studies, which often require the mechanical characterization of the TPP fabricated structures. To satisfy physiological requirements, most of the mechanical characterizations need to be conducted in liquid. However, previous characterizations of TPP fabricated structures are all conducted in air due to the limitation of conventional micro- and nanoscale mechanical testing methods. In this study, a new experimental method is reported for testing the mechanical properties of TPP-printed microfibers in liquid. The experiments show that the mechanical behaviors of the microfibers tested in liquid are significantly different from those tested in air. By controlling the TPP writing parameters, the mechanical properties of the microfibers can be tailored over a wide range to meet a variety of mechanobiology applications. In addition, it is found that, in water, the plasticly deformed microfibers can return to their predeformed shape after tensile strain is released. The shape recovery time is dependent on the size of microfibers. The experimental method represents a significant advancement in mechanical testing of TPP fabricated structures and may help release the full potential of TPP fabricated 3D tissue scaffolds for mechanobiological studies.     

On-line liquid backflush—a new technique for GC analysis of samples containing nonvolatile material

Deteriorated performance in capillary GC is often caused by the deposition of nonvolatile components in capillary GC inlets. In many cases this is a limiting factor in trace analysis of complex samples. Volatile solutes of the sample interact with the high boiling material, which leads to broadened and unsymmetrical peaks. This paper describes a novel technique which utilizes independent temperature programs of a retention gap and the column as well as a liquid backflush of the retention gap. Samples that normally contaminate the retention gap after a few injections can be repeatedly injected over extended periods of time, while the chromatographic resolution is retained.