一种健壮有效的决策树改进模型

提出了一种健壮有效的决策树改进模型AJD3.该决策树模型基于经典的ID3决策树模型,在属性的选取上进行了改进.利用引入属性优先关联度参数计算节点的修正信息增益,并选择具有最高修正增益的属性作为当前节点的测试属性.实验表明,AID3决策树模型在提高分类准确率的同时,有效地增强了模型的健壮性.     

Labview在BP神经网络谐波监测中的实现

为了减少电力系统中谐波的危害,本文将神经网络理论应用于谐波监测,使用界面设计方便美观的Labview编程语言制作了一种谐波监测仪。文中构造了一个目前应用最为广泛的BP网络,给出了网络算法,形成了训练样本。并用基于Labview工作平台的G语言,编写了训练程序,将所形成的样本对网络进行了训练,然后对训练好的网络进行仿真实验。结果表明:所设计的神经网络可很好的监测谐波,并对谐波参数的显示设计了很好的人机界面。     

Ultrafast Switchable Passive Radiative Cooling Smart Windows with Synergistic Optical Modulation

Switchable passive radiative cooling (PRC) smart windows can modulate sunlight transmission and spontaneously emit heat to outer space through atmospheric transparent window, presenting great potential in building energy conservation. However, realizing stable and on-demand control of the cooling efficiency for PRC materials is still challenging. Herein, an electro-controlled polymer-dispersed liquid crystal (PDLC) smart window showing PRC property is designed and prepared by adding mid-infrared emitting reactive monomers into the conventional PDLC matrix. It is found that not only the electro-optical properties but also the PRC efficiency of PRC PDLC film are tunable by regulating the content of the mid-infrared emitting components, film thickness, and micromorphology. This advanced PRC PDLC material achieves a near/sub-ambient temperature when the solar irradiance is below 400 W m⁻² and can dynamically manage daytime cooling efficiency. Importantly, its PRC efficiency is capable of being tuned in an on-demand and ultrafast millisecond-scale way, whose controllable transparency enables multistage heat regulation. This study is hoped to provide new inspiration in the preparation of advanced optical devices and energy-efficient equipment.     

Boosted Photocurrent via Heating BiFeo3 Thin Film for UV Photodetector at Wide Temperature Range

The current research on ferroelectric photovoltaic materials is concentrated on enhancing the output photocurrent. As solar cells operate at high temperatures, it is crucial to take into account the effect of increasing temperatures on ferroelectric photovoltaics. In this study, an LNO (lanthanum nickelate, LaNiO₃)/BFO (bismuth ferrate, BiFeO₃)/ITO (indium tin oxide) device is constructed on a mica substrate by sol–gel method. The device achieves output photocurrent enhancement at a wide temperature range (33–183 °C), with the largest photocurrent enhancement at 130 °C, which is 178% relative to room temperature, and the output power is also increased by 9.88 times. At the same time, compared with BFO bulk, it is found that the performance of BFO film is always higher than that of bulk in the test temperature range, and the output photocurrent of BFO film at room temperature is 104 times higher than that of bulk. This article investigates the effect of high temperatures on ferroelectric photovoltaics and also provides a strategy for enhancing the photovoltaic performance of ferroelectric films, providing guidance for future applications of ferroelectric films in flexible solar cells and other applications.     

Light-Responsive Programmable Shape-Memory Soft Actuator Based on Liquid Crystalline Polymer/Polyurethane Network

Liquid crystalline polymers (LCPs), especially liquid crystalline elastomers (LCEs) can generate ultrahigh shape change amplitude but has lower mechanical strength. Although some attempts have been tried to improve the mechanical performance of LCE, there are still limitations including complicated fabrication and high actuation temperature. Here, a versatile method is reported to fabricate light-driven actuator by covalently cross-linking polyurethane (PU) into LCP networks (PULCN). This new scheme is distinct from the previous interpenetrating network strategy, the hydrogen bonds and covalent bonds are used in this study to improve the miscibility of non-liquid-crystalline PU and LCP materials and enhance the stability of the composite system. This material not only possesses the shape memory properties of PU but shows shape-changing behavior of LCPs. With a shrinkage ratio of 20% at the phase transition temperature, the prepared materials reached a maximum mechanical strength of 20 MPa, higher than conventional LCP. Meanwhile, the resulting film shows diverse and programmable initial shapes by constructing crosslinking density gradient across the thickness of the film. By integration of PULCN with near-infrared light-responsive polydopamine, local and sequential light control is achieved. This study may provide a new route for the fabrication of programmable and mechanically robust light-driven soft actuator.     

基于深度森林的多级特征融合SAR目标识别

大多数传统的合成孔径雷达(SAR)目标识别方法仅仅使用了单一的幅度特征,但是由于斑点噪声的存在,仅仅使用幅度特征会限制识别的性能.为了进一步提高SAR目标识别的性能,该文提出了一个基于深度森林的多级特征融合SAR目标识别方法.首先,在特征提取阶段,提取了多级幅度特征和多级密集尺度不变特征变换(Dense-SIFT)特征.幅度特征反映了目标反射强度,Dense-SIFT特征描述了目标的结构特征.而多级特征可以从局部到全局表征目标.随后,为了更完整、充分地反映SAR目标信息,借鉴深度森林的思想对多级幅度特征和多级Dense-SIFT特征进行联合利用.一方面通过堆叠的方式不断将多级幅度特征和多级Dense-SIFT特征进行融合,另一方面通过逐层的特征变换挖掘深层信息.最后利用得到的深层融合特征对目标进行识别任务.该文在MSTAR数据集上进行对比实验,实验结果表明所提算法在性能方面取得了提升,且其性能对超参数设置具有一定的鲁棒性.     

基于MSP430F149的便携式超声波测距仪设计

根据超声波测距原理,设计了一种基于MSP430F149的便携式超声波测距仪;详细介绍了该测距仪硬件电路的设计,并给出了该测距仪软件主程序流程。测试结果表明,该测距仪运行稳定、性能可靠,测量精度可达2 mm。