Vehicle and pedestrian detection method based on improved YOLOv4-tiny

Aiming at the problem of low detection accuracy of vehicle and pedestrian detection models, this paper proposes an improved you only look once v4 (YOLOv4)-tiny vehicle and pedestrian target detection algorithm. Convolutional block attention module (CBAM) is introduced into cross stage partial Darknet-53 (CSPDarknet53)-tiny module to enhance feature extraction capabilities. In addition, the cross stage partial dense block layer (CSP-DBL) module is used to replace the original simple convolutional module superposition, which compensates for the high-resolution characteristic information and further improves the detection accuracy of the network. Finally, the test results on the BDD100K traffic dataset show that the mean average precision (mAP) value of the final network of the proposed method is 88.74%, and the detection speed reaches 63 frames per second (FPS), which improves the detection accuracy of the network and meets the real-time detection speed.     

Moist-Electric Generator with Efficient Output and Scalable Integration Based on Carbonized Polymer Dot and Liquid Metal Active Electrode

Moisture-enabled electricity generation (MEG) is highly promising in next-generation energy conversion. However, the practical applications of existing MEG devices are limited due to their low current and voltage outputs, strong dependence on high moisture, and inflexible nature. Herein, an efficient MEG integrated with flexible, all-weather, and scalable fabrication characteristics based on the rational combination of carbonized polymer dots (CPDs) and liquid metal (LM) active electrodes is developed for the first time. Remarkably, the fabricated MEG device can produce a stable voltage output of 800 mV and a record high current density of 1640 µA cm⁻². Even at a low air humidity of 15%, the MEG device can provide a high voltage output of 0.65 V and a considerable current density of 12 µA cm⁻². The prompted diffusion of hydrogen ions in CPDs and the additional metal ions ionized from the LM electrode contribute synergistically to the high electricity generation. Additionally, the device can be easily integrated on various flexible substrates and generate an ultrahigh voltage of 210 V to power commercial electronics, showing great potential in large-scale fabrication and application.     

基于插入WO3缓冲层并五苯有机场效应晶体管性能的研究

The pentacene-based organic field effect transistor （OFET） with a thin transition metal oxide （WO3） layer between pentacene and metal （AI） source/drain electrodes was fabricated. Compared with conventional OFET with only metal AI source/drain electrodes, the introduction of the WO3 buffer layer leads to the device performance enhancement. The effective field-effect mobility and threshold voltage are improved to 1.90 em2/（V.s） and 13 V, respectively. The performance improvements are attributed to the decrease of the interface energy barrier and the contact resistance. The results indicate that it is an effective approach to improve the OFET performance by using a WO3 buffer layer.     

基于阈值判断的动态源路由协议算法优化设计

针对移动自组织网动态源路由协议(DSR)中旧路径中断至新路径建立期间存在的丢包和端到端时延问题,设计了一种基于阈值判断的路由发现和维护算法,以优化路由协议性能。该方法首先定义节点电池能量和节点接收信号能量双阈值等级,分别用于衡量节点和既存链路状态;然后修改路由发现和路由维护算法,使动态源路由具备阈值判断功能。分别对增强型动态源路由协议(EDSR)包传递率和端到端时延等性能随节点数目和移动速度等参数的变化规律进行了软件仿真。实验结果表明,节点数目超过150或者速度小于10 m/s时,EDSR端到端时延和包传递率优于DSR,其他性能基本相当,因此EDSR协议性能得到明显改善,为低速大规模移动自组网路由协议提供了重要参考。     

基于图像运动检测的子场编码优化算法

针对等离子显示器中运动图像产生的动态伪轮廓现象,提出了一种基于图像运动检测的子场编码优化算法。该算法在分析常用动态伪轮廓和运动补偿算法的基础上,通过检测图像数据的变化差值判断图像的运动状态,并通过图像运动状态选择不同的显示灰度构成方式:静态图像选择全灰度级编码,动态图像选择较少灰度级编码,通过优化编码方式消除图像的动态伪轮廓。实验表明,算法减少了不同运动速度的图像因算法引起的人工纹理,保证了静态图像的显示细节,达到改善显示图像画质的目的。此外,该算法还具有处理速度快,易于硬件实现等优点。     

A Biodegradable Antigen Nanocapsule Promotes Anti-Tumor Immunity via the cGAS-STING Pathway

Materials-based antigen delivery systems can augment the immune response by improving antigen uptake in antigen-presenting cells, targeting lymph nodes, prolonging antigen exposure, enhancing cross-presentation, etc. Recent research revealed that some antigen carriers activate the innate immune pathways without additional adjuvant. Here, a vaccine delivery platform (antigen nanocapsules) constructed by a one-step in situ polymerization is reported, weaving a biodegradable polymer network around the antigen surface. This simple technology allowed us to study the immunomodulatory effect of various antigen carriers. An antigen nanocapsule (NC7) capable of inducing dendritic cell activation and cross-presentation is identified. Further mechanistic studies revealed that NC7 activated the cGAS-STING pathway in a cGAS-dependent manner. Moreover, the subcutaneously injected NC7 accumulated in the lymph nodes and elicited strong cytotoxic T cell immunity and T cell memory against established cancer. Collectively, the often-neglected immunomodulatory effect of various cationic antigen carriers, enabling potential application in cancer vaccines, is uncovered.     

A Low-Hysteresis and Highly Stretchable Ionogel Enabled by Well Dispersed Slidable Cross-Linker for Rapid Human-Machine Interaction

Ionic conductive soft materials for mimicking human skin are a promising topic since they can be thought of as a possible basis for biomimetic sensing. In pursuit of devices with a long working range and low signal delay, conductive materials with low hysteresis and good stretchability are highly demanded. To overcome the challenges of highly stretchable conductive materials with good resilience, herein a chemical design is proposed where polyrotaxanes act as topological cross-linkers to enhance the stretchability by sliding-induced reduced stress concentration while the compatible ionic liquid is introduced as a dispersant for low hysteresis. The obtained ionogels exhibit versatile properties more than low hysteresis (residual strain = 7%) and good stretchability (550%), and also anti-fatigue, biocompatibility, and good adhesion. The low hysteresis is attributed to lower energy dissipation from the well-dispersed polyrotaxanes by compatible ionic liquids. The mechanism provides a new insight in fabricating highly stretchable and low-hysteresis slide-ring materials. Furthermore, the conductivity of the ionogels and their responses to strains and temperatures are measured. Benefiting from the good conductivity and low hysteresis, the ionogel is applied to develop a wireless communication system to realize rapid human-machine interactions.     

Optical Design of Silica Aerogels for On-Demand Thermal Management

Silica aerogels, a type of porous material featuring extra low density and thermal conductivity, have drawn increasing interest from both academia and industry owing to their excellent thermal insulation performance. However, thermal insulation is always the single consideration when silica aerogels are used for thermal management. In this study, the on-demand thermal management (ODTM) of silica aerogel with either passive thermal insulation, passive heating, or passive cooling in different environments is revealed. The ODTM behavior of silica aerogels can be simply fulfilled through their optical property variations such as solar light transparency and infrared emissivity, which are controllable via the microstructures of the building blocks and surface composition design. Robust solar heating of 25 °C higher than the ambient in the daytime and sub-ambient cooling of 7 °C at night is achieved with the traditional transparent silica aerogel. Interestingly, sub-ambient cooling of 5 °C in the daytime and a warmer state on cold nights is achieved by modifying its solar transmittance and infrared emissivity. This study guides a comprehensive understanding of the thermal management behavior of silica aerogels and leads to ODTM applications of silica aerogels by tailoring their optical and thermal conductivity properties.     

基于场景的扫描线非均匀性校正算法

目前扫描红外热像仪有着广泛的应用,且一般采用基于定标的非均匀性校正方法;该类方法不能有效地校正图像中出现的扫描线。针对这一问题,本文对扫描线的表现进行分析,提出了一种基于场景的非均匀性校正算法,该算法采用图像像素点邻域标准差作为基础,通过类标准差对图像的差异信息进行计算,能够排除复杂场景的干扰,成功地识别扫描线,并对扫描线提取补偿。该算法计算简单,可以根据场景进行自适应校正,校正效果显著,具有一定的工程应用价值。