直通链路技术的发展与展望

直通链路技术己广泛应用于车联网场景。对于直通链路技术的潜在技术方向给出可行的建议，包括传统直通链路技术的增强方向，如载波聚合、使用非授权频谱等；侧行链路对于中继场景的应用扩展，包括终端到终端之间的中继，以及中继的多链接场景；在高精度定位场景使用直通链路技术。并且，给出直通链路技术与各种新技术的融合应用，如智能反射面与区块链技术，从而解决直通链路技术自身的缺陷。     

基于特征建模的TDD系统大气波导干扰控制体系研究

大气波导干扰是特定气象条件下发生的时分双工（time-division duplex，TDD）系统内干扰，是TDD移动通信系统大规模组网面临的顽疾。在总结分析大气波导干扰成因和分类等的基础上，对大气波导干扰进行建模和表征，验证了海量干扰源在时域和频域的功率集总特征，并结合大量4G/5G现网实测数据给出了典型条件下内陆波导和海面波导的量化干扰信号传播模型，对于干扰的预测和预防具有重要意义。基于干扰特征，给出了TDD系统预防大气波导干扰的帧结构与组网的4项设计原则，5G现网数据表明干扰控制方案有效，上行干扰下降10 dB以上，相关原则对于6G系统的设计也具有指导意义。     

外军地面无人装备发展现状与趋势研究

随着通信、计算机、人工智能和传感器等技术的飞速发展与广泛应用，地面无人装备日趋活跃在作战、搜救和消防等多种军民用任务场景中。对于军事领域而言，地面无人装备是陆军和其他地面作战部队迈向无人化、智能化战场的核心装备，将在未来陆战场各类作战任务中发挥不可替代的作用。通过对外军各类典型地面无人装备发展现状进行深入分析，总结出该装备领域的发展特点与趋势，对地面无人装备发展有一定的借鉴意义。     

Lighting the Path: Light Delivery Strategies to Activate Photoresponsive Biomaterials In Vivo

Photoresponsive biomaterials are experiencing a transition from in vitro models to in vivo demonstrations that point toward clinical translation. Dynamic hydrogels for cell encapsulation, light-responsive carriers for controlled drug delivery, and nanomaterials containing photosensitizers for photodynamic therapy are relevant examples. Nonetheless, the step to the clinic largely depends on their combination with technologies to bring light into the body. This review highlights the challenge of photoactivation in vivo, and presents strategies for light management that can be adopted for this purpose. The authors’ focus is on technologies that are materials-driven, particularly upconversion nanoparticles that assist in “direct path” light delivery through tissue, and optical waveguides that “clear the path” between external light source and in vivo target. The authors’ intention is to assist the photoresponsive biomaterials community transition toward medical technologies by presenting light delivery concepts that can be integrated with the photoresponsive targets. The authors also aim to stimulate further innovation in materials-based light delivery platforms by highlighting needs and opportunities for in vivo photoactivation of biomaterials.     

基于1D CNN与XGBoost的恶意代码纹理检测

恶意代码数量已经呈现爆炸式增长,对于恶意代码的检测防护显得尤为重要.近几年,基于深度学习的恶意代码检测方法开始出现,基于此,提出一种新的检测方法,将恶意代码二进制文件转化为十进制数组,并利用一维卷积神经网络(1 Dimention Convolutional Neural Networks,1D CNN)对数组进行分类和识别.针对代码家族之间数量不平衡的现象,该算法选择在分类预测上表现良好的XGBoost,并对Vision Research Lab中的25个不同恶意软件家族的9458个恶意软件样本进行了实验.实验结果表明,所提的方法分类预测精度达到了97％.