基于太阳黑子群数据的多模态太阳耀斑预报模型

当前的太阳耀斑预报模型主要通过统计关系建立,直接将从太阳黑子群磁图中提取的特征参量作为模型输入,系统的自主性低,导致图像数据中包含的与太阳耀斑相关的高阶抽象信息难以被充分利用,进而限制模型预报的精度。为解决当前太阳耀斑预报中数据利用不充分的问题,文章将海量太阳观测数据与先进的人工智能技术结合,综合利用太阳活动区磁场观测图、磁场特征参量和对应的耀斑事件标签,并借助全连接神经网络高精确率以及卷积神经网络高召回率和可有效提取高层语义信息的优点,构建基于深度学习的多模态太阳耀斑预报模型。实验证明该模型的主要评价指标结果比其他模型至少提高7.8%。     

基于无线传感器网络的低功耗目标探测系统设计与实现

基于无线传感器网络的目标探测系统利用磁场、振动、声音等环境参数的变化实现对移动目标的探测。针对无线传感器网络的能量受限问题,从传感器节点硬件选型、节点休眠设计、网络拓扑结构设计和数据传输协议设计等方面出发,进行系统低功耗设计优化。实验结果表明,该探测系统在传感器节点采用容量4 A·h锂电池,采样率10 Hz情况下,平均工作电流仅为3.78 mA,连续工作时长可达33天。     

迭代法求解航天器热网络方程的收敛条件与时间步长优化

热分析计算是航天器热设计验证和在轨温度预示的重要手段,其实质是对描述航天器在轨状态的能量平衡方程即热网络方程的求解。作为典型的非线性问题,航天器热网络方程通常采用迭代法求解,且目前尚无完善的理论可以判断迭代过程的收敛性和收敛条件。文章根据航天器热控设计的工程实际,提出一种可行的近似方法,实现了热网络方程的线性化;在此基础上对迭代法求解过程中的收敛性进行分析,确定了收敛条件下对节点热参数和时间步长的约束关系;并进一步分析时间步长与计算效率的关系,提出了以计算效率为优化目标的最优时间步长确定方法。最后通过某典型航天器热物理模型的计算验证了上述方法的正确性。     

基于神经网络的磁平静期赤道电集流预测

利用BP神经网络技术分别对2008年后磁平静期印度扇区、秘鲁扇区以及CHAMP卫星的赤道电集流（EEJ）变化进行预测,其中神经网络训练数据为对应的2000—2007年磁平静期EEJ观测数据,输入参量为天数、地方时、太阳天顶角、太阳活动指数（F_10.7）、太阴时以及卫星地理经度,输出参量为EEJ.对EEJ预测结果进行了统计学分析,并且与实际观测结果进行对比.结果表明：BP神经网络对事件中EEJ的变化具有很好的预测能力,预测结果能够反映EEJ的重要分布特征;EEJ预测值与观测值之间具有很好的相关性,其中地磁台站观测值与预测值相关性系数可达85%以上.此外,将BP神经网络模型的预测结果与Yamazaki提出的经验模型结果进行对比,结果显示BP神经网络与其经验模型性能相当.研究结果表明,BP神经网络技术在平静期EEJ变化预测方面性能优异,具有良好的应用前景.     

直流电机网络控制系统仿真

对直流电机网络闭环控制系统进行仿真分析,根据电机网络闭环控制系统原理建立仿真模型。对具有数据丢包的直流电机网络伺服控制系统输出响应特性进行仿真,比较系统分别采用PI控制及增益,可调PI控制方式时的输出特性。结果表明,相比常规PI控制增益可调PI控制方法在网络环境下具有更好的控制特性。该研究为网络伺服控制技术提供可靠的参考依据。     

Exploiting global information in complex network repair processes

Robustness of complex networks has been studied for decades,with a particular focus on network attack.Research on network repair,on the other hand,has been conducted only very lately,given the even higher complexity and absence of an effective evaluation metric.A recently proposed network repair strategy is self-healing,which aims to repair networks for larger compo nents at a low cost only with local information.In this paper,we discuss the effectiveness and effi ciency of self-healing,which limits network repair to be a multi-objective optimization problem and makes it difficult to measure its optimality.This leads us to a new network repair evaluation metric.Since the time complexity of the computation is very high,we devise a greedy ranking strategy.Evaluations on both real-world and random networks show the effectiveness of our new metric and repair strategy.Our study contributes to optimal network repair algorithms and provides a gold standard for future studies on network repair.     

Air route network optimization in fragmented airspace based on cellular automata

Air route network optimization,one of the essential parts of the airspace planning,is an effective way to optimize airspace resources,increase airspace capacity,and alleviate air traffic con gestion.However,little has been done on the optimization of air route network in the fragmented airspace caused by prohibited,restricted,and dangerous areas (PRDs).In this paper,an air route network optimization model is developed with the total operational cost as the objective function while airspace restriction,air route network capacity,and non-straight-line factors (NSLF) are taken as major constraints.A square grid cellular space,Moore neighbors,a fixed boundary,together with a set of rules for solving the route network optimization model are designed based on cellular automata.The empirical traffic of airports with the largest traffic volume in each of the 9 flight information regions in mainland China is collected as the origin-destination (OD) air port pair demands.Based on traffic patterns,the model generates 35 air routes which successfully avoids 144 PRDs.Compared with the current air route network structure,the number of nodes decreases by 41.67％,while the total length of flight segments and air routes drop by 32.03％ and 5.82％ respectively.The NSLF decreases by 5.82％ with changes in the total length of the air route network.More importantly,the total operational cost of the whole network decreases by 6.22％.The computational results show the potential benefits of the model and the advantage of the algorithm.Optimization of air route network can significantly reduce operational cost while ensuring operation safety.     

基于自适应径向基网络的舰船RCS统计特征识别方法

文章采用舰船RCS频域起伏序列的均值、标准差为识别特征向量,利用提出的基于样本密度的自适应径向基网络,进行舰船分类识别研究。自适应径向基网络采用改进的自适应PSO方法估计样本密度最优邻域半径,实现径向基网络中心的自适应选择。改进的自适应PSO方法采用能反映样本聚类特点的BWP指标为适应度评价函数,采用快慢结合的高斯自适应惯性权重调节策略,提高了最优样本密度邻域半径的搜索速度和精度。实验结果表明,自适应径向基网络能自适应获得径向基网络最优识别率对应的RBF中心及其位置分布,减少了对建模人员经验的依赖,提高了反舰导弹对舰船类型的识别分类能力。     

A regional tropospheric model for airborne GPS applications

A regional tropospheric model can be constructed using surveys from GPS ground networks. Using this model the tropospheric delays of a kinematic station within the region can be interpolated. However, such a model is generally not suitable for an airborne platform high above the ground networks. In this paper, a method of constructing a regional tropospheric model for airborne GPS applications is described. First, the kinematic station in the air is projected onto the ground. Then the tropospheric delays at projected point are interpolated from those of the ground networks. Finally, the tropospheric delays at projected point are extended upward to the airborne platform using pressure and temperature gradients and humidity exponential function. For validation of this method, the data of airborne campaign carried out by BKG in cooperation with GFZ and BGR 2008 in the northern Alps are used. The results show that GPS kinematic positioning precision in height component can be improved using this method.     

基于人工神经网络的停机坪安全预警研究

停机坪是机场的重要组成部分,停机坪安全是机场安全关注的重点。影响停机坪安全作业的要素有很多,通过对这些要素进行监控和预警,可以减少停机坪的安全隐患,保证机场正常运行。根据济南遥墙机场的主要数据,运用因子分析法对影响停机坪安全的主要因素进行分类,提取公共因子。并运用人工神经网络（ANN）方法,对这些公共因子进行神经网络训练和预测分析,最终得到遥墙机场停机坪的安全情况。