基于数据质量评价的目标估计

针对目标估计过程需要大量人工参与、自动化程度低的问题，提出了基于数据质量评价的目标估计方法。利用目标数据质量评价方法，对不同传感器得到的目标数据质量进行科学、有效的测度和评价，并根据质量得分动态调整各数据源在目标估计过程中所占的权重，从而减少人工干预，提高目标估计效能。仿真试验结果证明了该方法的有效性。     

基于卷积神经网络的5G蜂窝网络无线定位方法

5G蜂窝网络发展迅猛，其覆盖面积将逐渐增大，因此使用5G蜂窝网络进行定位是有研究潜力的研究方向。本文提出一种新的深度学习技术来实现高效、高精度和低占用的定位，以代替传统指纹定位过程中繁重的指纹库生成以及距离计算。该方法建立了一个特殊的卷积神经网络，并根据5G天线信号的接收信号强度指示、相位和到达角等特征量，选择合适的输入数据格式构造样本组建训练集，对该卷积神经网络进行训练。训练得到的卷积神经网络可以替代指纹定位中的庞大指纹库，非常有利于直接在5G移动设备端实现定位。虽然卷积神经网络在训练过程中需要大量时间，但在训练完毕后直接进行分类定位的速度非常快，可以保障定位实现的实时性。本文所实现的卷积神经网络权重与偏置所占内存不到0.5 MB，且能够在实际应用环境中以95%的定位准确率以及0.1 m的平均定位精度实现高精度定位。     

气候变化背景下滇中引水工程水源区与受水区 降水丰枯遭遇分析

基于历史实测降水数据与全球气候模型预估数据，使用 Morlet 小波方法分析滇中引水工程水源区与受水区 降水序列的周期变化和未来的降水趋势。同时，采用 Copula 函数计算历史时期（1960—2021 年）与未来时期 （2022—2100 年）水源区与受水区降水丰枯异步或丰枯同步的概率。结果表明：1960—2021 年降水序列存在 26~39?a、18~25?a、4~7?a 的 3 类时间尺度的周期变化，2022—2100 年降水序列存在 38~55?a、18~30?a、5~12?a 的 3 类时间尺度的周期变化，降水量呈现“多—少—多”的循环交替，预计未来 10~20?a 将持续处于降水较多的时期； 过去 62?a，水源区和受水区降水丰枯异步频率 36.4%，同期丰水年频率为 25.3%，同期枯水年频率小于 30%，水源区 和受水区具有水量互补的引水条件，两区域之间存在着水量补偿特征；与历史丰枯遭遇对比，未来降水量丰枯同 步频率均呈现减小的趋势，丰枯异步呈现增加的趋势，同枯和源枯受丰的频率减少，未来有利于调水的降水丰枯 组合概率平均增加 3.75%；在近、中、远期预估中，从 SSP1-2.6 情景过渡到 SSP5-8.5 情景，SSP5-8.5 情景下降水量 丰枯异步频率比 SSP1-2.6 情景大，说明水源区与受水区的降水区域差异变大，降水时空差异更加显著。通过对滇 中引水工程水源区与受水区降水量丰枯遭遇的综合分析、定量评估和模拟预测，为滇中引水工程水资源调度协同 一体化提供数据支撑及参考依据。     

Dynamic voltage scaling based energy-minimized partial task offloading in fog networks

Wireless Networks - With the dynamic voltage scaling (DVS) technology, the terminal node (TN) can dynamically adjust its computational speed, thus providing a new way to save energy during task...     

Multi-UAV Collaborative Wireless Communication Networks for Single Cell Edge Users

Mobile Networks and Applications - Due to fast deployment, strong survivability and flexible scheduling, multiple unmanned aerial vehicles (UAV) can synergistically optimize the trajectory and...     

Graph-Theoretic Method on Topology Identification of Stochastic Multi-weighted Complex Networks with Time-Varying Delayed Coupling Based on Adaptive Synchronization

Neural Processing Letters - This paper is concerned with topology identification of stochastic multi-weighted complex networks based on adaptive synchronization. Different from previous work,...     

Effects of structural parameters on water film properties of transpiring wall reactor

Corrosion and salt deposition problems severely restrict the industrialization of supercritical water oxidation. Transpiring wall reactor can effectively weaken these two problems by a protective water film. In this work, methanol was selected as organic matter, and the influences of vital structural parameters on water film properties and organic matter removal were studied via numerical simulation. The results indicate that higher than 99% of methanol conversion could be obtained and hardly affected by transpiration water layer, transpiring wall porosity and inner diameter. Increasing layer and porosity reduced reactor center temperature, but inner diameter's influence was lower relatively. Water film temperature reduced but coverage rate raised as layer, porosity, and inner diameter increased. Notably, the whole reactor was in supercritical state and coverage rate was only approximately 85% in the case of one layer. Increasing reactor length affected slightly the volume of the upper supercritical zone but enlarged the subcritical zone.