基于深度强化学习的无人机栖落机动控制策略设计

无人机栖落机动飞行是一种无需跑道的降落方法,能够提升无人机在复杂环境下执行任务的适应能力。针对具有高非线性、多约束特性的无人机栖落机动过程,提出了一种基于模仿深度强化学习的控制策略设计方法。首先,建立了固定翼无人机栖落机动的纵向非线性动力学模型,并设计了无人机栖落机动的强化学习环境。其次,针对栖落机动状态动作空间大的特点,为了提高探索效率,通过模仿专家经验的方法对系统进行预训练。然后,以模仿学习得到的权重为基础,采用近端策略优化方法学习构建无人机栖落机动的神经网络控制器。最后,通过仿真验证了上述控制策略设计方法的有效性。     

基于卡尔曼滤波器及神经网络的发动机故障诊断

提出了一种基于卡尔曼滤波器及神经网络的航空燃气涡轮发动机气路故障诊断的方法.该方法用卡尔曼滤波器来估计发动机可测参数的变化量,再由神经网络来映射发动机性能参数的变化量,并据此进行发动机气路故障诊断.数字仿真表明,该方法是可行的,有效的.      

Load balancing strategy and its lookup-table enhancement in deterministic space delay/disruption tolerant networks

Space networks, in which connectivity is deterministic and intermittent, can be modeled by delay/disruption tolerant networks. In space delay/disruption tolerant networks, a packet is usually transmitted from the source node to the destination node indirectly via a series of relay nodes. If anyone of the nodes in the path becomes congested, the packet will be dropped due to buffer overflow. One of the main reasons behind congestion is the unbalanced network traffic distribution. We propose a load balancing strategy which takes the congestion status of both the local node and relay nodes into account. The congestion status, together with the end-to-end delay, is used in the routing selection. A lookup-table enhancement is also proposed. The off-line computation and the on-line adjustment are combined together to make a more precise estimate of the end-to-end delay while at the same time reducing the onboard computation. Simulation results show that the proposed strategy helps to distribute network traffic more evenly and therefore reduces the packet drop ratio. In addition, the average delay is also decreased in most cases. The lookup-table enhancement provides a compromise between the need for better communication performance and the desire for less onboard computation.     

Virtual sensing method for monitoring vibration of continuously variable configuration structures using long short-term memory networks

Vibration monitoring by virtual sensing methods has been well developed for linear time-invariant structures with limited sensors. However, few methods are proposed for Time-Varying (TV) structures which are inevitable in aerospace engineering. The core of vibration monitoring for TV structures is to describe the TV structural dynamic characteristics with accuracy and efficiency. This paper propose a new method using the Long Short-Term Memory (LSTM) networks for Continuously Variable Configuration Structures (CVCSs), which is an important subclass of TV structures. The configuration parameters are used to represent the time-varying dynamic characteristics by the “freezing” method. The relationship between TV dynamic characteristics and vibration responses is established by LSTM, and can be generalized to estimate the responses with unknown TV processes benefiting from the time translation invariance of LSTM. A numerical example and a liquid-filled pipe experiment are used to test the performance of the proposed method. The results demonstrate that the proposed method can accurately estimate the unmeasured responses for CVCSs to reveal the actual characteristics in time-domain and modal-domain. Besides, the average one-step estimation time of responses is less than the sampling interval. Thus, the proposed method is promising to on-line estimate the important responses of TV structures.     

基于模拟退火算法预测储层参数

讨论了如何将地震资料、测井信息和先验地质知识有机地结合起来预测储层参数，以便揭示储层内部细节和纵横向变化特征。通过将模拟退火算法与人工神经网络相结合，能够解决非线性全局优化问题，勾画出地震信息和储层参数之间的复杂关系，为储层特征描述和储层横向预测提供提供直观、可靠的地层参数剖面一孔隙度、砂泥质百分含量、含水饱和度及渗透率等剖面。     

分布交互仿真的一个原型系统—DIFTS

介绍了分布交互仿真的概念及其背景，分析了它的特点及涉及的关键技术，然后，设计了分布交互仿真的一个原型系统－ＤＩＦＴＳ，分析了ＤＩＦＴＳ的功能模型，并对其进行了功能分解，给出了各模块的功能，最后，介绍了系统的硬件组成及支持工具。     

具有神经元网络补偿的隐式广义预测控制的硬件设计与分析

针对隐式广义预测控制算法在被控系统复杂性增加时有效性降低、实时性变差的问题，本文将神经元网络作为该算法的补偿，提出了一种具有神经元网络补偿的隐式广义预测控制方法，并着重讨论了神经元网络的硬件设计和ＮＩＧＰＣ的实现。仿真结果表明：ＮＩＧＰＣ不仅保留了ＩＧＰＣ鲁棒性强等特点，更主要的是它提高了ＩＧＰＣ的的实时性，并使之具有良好的跟踪特性。     

基于神经网络的多模态故障检测

研究了使用RBF网络进行多模态辨识的结构和其在多模态故障检测中的应用。在信号满足持续激励的条件下 ,状态估计误差渐进趋向于零 ,辨识器的参数估计能够达到满意的效果。使用多个神经网络辨识器 ,对一非线性系统的多模态故障检测进行了仿真研究     

STABILITY OF THE HIGH-ORDER BAM AND ITS PERFORMANCE ANALYSIS

由Ｔａｉ提出的高阶双向联想存储器（ＨＯＢＡＭ）推广了Ｋｏｓｋｏ的ＢＡＭ，它不仅拥有ＢＡＭ的优点，而且放宽了连续性假定以得到可靠的回忆，并大大改进了ＢＡＭ的存储容量和纠错性能。然而，Ｔａｉ对于ＨＯＢＡＭ的性能分析仅限于计算机模拟，并没有给出理论分析结果。本文填写了此方面空白，并给出ＨＯＢＡＭ稳定性和存储容量的理论证明，使得该系统能从理论上得以保证所有训练模式对成为其稳定点。     

Information space of sensor networks: Lagrangian,energy-momentum tensor,and applications

It is a challenge to investigate the interrelationship between the geometric structure and performance of sensor networks due to the increasingly complex and diverse architecture of them. This paper presents two new formulations for the information space of sensor networks, including Lagrangian and energy–momentum tensor, which are expected to integrate sensor networks target tracking and performance evaluation from a unified perspective. The proposed method presents two geometric objects to represent the dynamic state and manifold structure of the information space of sensor networks. Based on that, the authors conduct the property analysis and target tracking of sensor networks. To the best of our knowledge, it is the first time to investigate and analyze the information energy–momentum tensor of sensor networks and evaluate the performance of sensor networks in the context of target tracking. Simulations and examples confirm the competitive performance of the proposed method.