基于BP神经网络的GNSS-R海面风速反演

针对GNSS-R进行海面风速反演过程中时频域相关物理量较多，数据耦合性强等问题，提出了基于反向传播（BP）神经网络反演海面风速的方法。建立反演过程中相关观测量与风速的对应关系，选取多观测量作为输入，对输入数据进行处理，设置神经元与激励函数，使用BP神经网络自适应调整拟合参数，将风速作为神经网络输出端的特征量提取。反演结果，风速≤ 20 m/s时，反演均方根误差RMSE=1.21 m/s，风速>20 m/s时反演均方根误差RMSE=2.54 m/s，反演结果优于使用时延相关曲线前沿斜率（LES）和时延多普勒相关功率均值（DDMA）方法得到的反演结果，且迭代次数较少，复杂度较低，证明该方法可以应用于GNSS-R海面风速反演。     

基于神经网络的飞行科目模板样本集建立方法

为建立某一飞行科目的模板样本集,提出一套基于神经网络和卡尔曼滤波的数据处理方法。任选一次试飞样本,建立Kohonen自组织神经网络进行参数降维、聚类分析、特征提取等,使样本量缩减90%以上,得到该科目的模板样本集。用处理后的样本训练BP神经网络,对其他未经处理的试飞样本进行载荷预测,误差均在3%之内,说明处理后的样本能代表该科目的数据特点,即为模板样本集。方法可以为飞机载荷监控数据库的完善工作服务。     

人工智能在OODA 循环中的应用

当前人工智能(AI) 是商业和工业，甚至军事应用领域中最热门的流行语。军事国防领域也已经将人 工智能发展成为了一种技术，并应用到军事作战和决策系统中。最初OODA 循环是最知名的空战战术模型， 后来发展成为各国空军的作战方法。OODA 循环对于任务的定义和决策过程的理解起到了重要的作用。本文主 要研究了人工智能在OODA 循环决策中的应用。讨论了未来OODA 循环各阶段中所需的AI 作用和技术，包 括可解释的人工智能和基于学习的人工智能。     

Robustness evaluation method for unmanned aerial vehicle swarms based on complex network theory

Unmanned Aerial Vehicle (UAV) swarms have been foreseen to play an important role in military applications in the future, wherein they will be frequently subjected to different disturbances and destructions such as attacks and equipment faults. Therefore, a sophisticated robustness evaluation mechanism is of considerable importance for the reliable functioning of the UAV swarms. However, their complex characteristics and irregular dynamic evolution make them extremely challenging and uncertain to evaluate the robustness of such a system. In this paper, a complex network theory-based robustness evaluation method for a UAV swarming system is proposed. This method takes into account the dynamic evolution of UAV swarms, including dynamic reconfiguration and information correlation. The paper analyzes and models the aforementioned dynamic evolution and establishes a comprehensive robustness metric and two evaluation strategies. The robustness evaluation method and algorithms considering dynamic reconfiguration and information correlation are developed. Finally, the validity of the proposed method is verified by conducting a case study analysis. The results can further provide some guidance and reference for the robust design, mission planning and decision-making of UAV swarms.     

适应拓扑变化的拥塞最小化网络更新策略

软件定义网络的出现推动了航空集群机载网络的发展，同时也为机载网络的一致性更新带来了挑战。针对拓扑变化导致的网络更新期间可用路径失效问题，提出适应拓扑变化的拥塞最小化网络更新策略。首先，为减少可用路径失效带来的数据包丢失，在更新之前采用重路由机制进行处理；然后，针对重路由机制可能引起的网络拥塞，设计了贪婪流迁移算法缓解拥塞；最后，通过瞬时拥塞最小化更新算法完成整个网络更新。仿真结果表明：与传统的拥塞一致性更新算法相比，所提更新策略用规则开销、瞬时网络拥塞方面的少量提高换取了网络更新期间数据包丢失的显著降低。     

基于可靠性分析的立方星网络维护架构优化

立方星编队或星座构成分布式空间传感器网络，可提高立方星执行复杂空间任务的能力。然而立方星易发生故障，其故障时间不确定性也导致了传感器网络性能的不稳定，这凸显了对立方星网络进行在轨维护的重要性。考虑立方星传感器网络的功能维持问题，描述了一种网络维护架构，通过定期发射、在轨备份立方星以及时更换故障立方星，从而提高网络对单星随机失效事件的快速响应与恢复能力。建立了该架构的运行成本模型，包括固定成本、储存成本和短缺成本。收集整理了真实的立方星寿命数据，并使用最大化拟合优度参数估计方法得到最优立方星寿命的随机模型。采用基于蒙特卡罗仿真的遗传算法优化备用立方星的补给时刻和补给数量，在备份成本与系统性能下降所带来的损失之间进行权衡，使得系统的综合收益最优。     

Evolution analysis of a UAV real-time operating system from a network perspective

With the flourishing development of Unmanned Aerial Vehicles (UAVs), the mission tasks of UAVs have become more and more complex. Consequently, a Real-Time Operating System (RTOS) that provides operating environments for various mission services on these UAVs has become crucial, which leads to the necessity of having a deep understanding of an RTOS. In this paper, an empirical study is conducted on FreeRTOS, a commonly used RTOS for UAVs, from a complex network perspective. A total of 85 releases of FreeRTOS, from V2.4.2 to V10.0.0, are modeled as directed networks, in which the nodes represent functions and the edges denote function calls. It is found that the size of the FreeRTOS network has grown almost linearly with the evolution of the versions, while its main core has evolved steadily. In addition, a k-core analysis-based metric is proposed to identify major functionality changes of FreeRTOS during its evolution. The result shows that the identified versions are consistent with the version change logs. Finally, it is found that the clustering coefficient of the Linux OS scheduler is larger than that of the FreeRTOS scheduler. In conclusion, the empirical results provide useful guidance for developers and users of UAV RTOSs.     

Crop parameters estimation by fuzzy inference system using X-band scatterometer data

Learning fuzzy rule based systems with microwave remote sensing can lead to very useful applications in solving several problems in the field of agriculture. Fuzzy logic provides a simple way to arrive at a definite conclusion based upon imprecise, ambiguous, vague, noisy or missing input information. In the present paper, a subtractive based fuzzy inference system is introduced to estimate the potato crop parameters like biomass, leaf area index, plant height and soil moisture. Scattering coefficient for HH- and VV-polarizations were used as an input in the Fuzzy network. The plant height, biomass, and leaf area index of potato crop and soil moisture measured at its various growth stages were used as the target variables during the training and validation of the network. The estimated values of crop/soil parameters by this methodology are much closer to the experimental values. The present work confirms the estimation abilities of fuzzy subtractive clustering in potato crop parameters estimation. This technique may be useful for the other crops cultivated over regional or continental level.     

基于神经网络的精确末制导律研究

针对动能杀伤武器重量极小化问题，基于神经网络技术、动态逆补偿法和非线性优化法，研究了两种精确末制导律，一种采用FFBP神经网络学习一要制导方程动态逆实现精神制导，另一种采用FFBP神经网络学习非线性优化解实现精确制导，仿真分析表明，前者对目标加速度估计误差过于敏感，无法实现；而后者在目标加速度有误差时仍能实现精确制导，估于偏置比例导引，并可以使动能杀伤武器重量极小化。     

基于多层前向神经网络的模拟电路故障诊断

文中提出把一个神经网络应用到电路故障诊断的故障字典法中,以提高故障诊断的适应能力。这样,就使得字典法不仅适用于电路的故障情况,还提高了故障诊断的有效性。模拟结果显示这种方法是有效的。