基于Q-Learning的深度神经网络自适应退避策略

针对无人机自组织网络,结合Q-Learning和深度神经网络,提出一种自适应退避策略,以提高基于竞争的MAC协议通信性能.以Matlab为仿真平台,仿真比较了自适应退避策略与二进制指数退避策略的性能.     

Surface morphology characterization of unidirectional carbon fibre reinforced plastic machined by peripheral milling

This paper aims to characterise surface morphology and 3D roughness parameters of unidirectional carbon fibre reinforced plastic（UD-CFRP） milled at 0°, 45°, 90°, and 135° fibre orientation angles（FOAs）. Side milling experiments are conducted on UD-CFRP laminates. Surface damage forms and texture direction of milled surface are analysed. Spatial frequency of defects on CFRP surface is quantitatively determined using radially averaged 2D PSD. The kinematicdynamic surface topography is reconstructe...     

Sequential dynamic resource allocation in multi-beam satellite systems:A learning-based optimization method

Multi-beam antenna and beam hopping technologies are an effective solution for scarce satellite frequency resources. One of the primary challenges accompanying with Multi-Beam Satellites(MBS) is an efficient Dynamic Resource Allocation(DRA) strategy. This paper presents a learning-based Hybrid-Action Deep Q-Network(HADQN) algorithm to address the sequential decision-making optimization problem in DRA. By using a parameterized hybrid action space,HADQN makes it possible to schedule the beam patte...     

Conceptual design and analysis of legged landers with orientation capability

The lander with orientation capability is of fundamental importance for the “Returning” mission because its body can be the base with a suitable launching angle for the ascender lifting off. This paper proposes the conceptual design and verification of Landers with Orientation Capability (LOCs). Firstly, the topological composition of the lander is analyzed and expressed in the form of Lie group, which provides the fundamental principle for the type synthesis of LOCs. Then, the type synthesis approach for designing LOCs is proposed, which includes designing the Equivalent Parallel Mechanism (EPM) and the auxiliary limbs. Numerous mechanisms of LOCs with two-dimensional rotational motions are obtained based on the type synthesis approach by developing and combining the EPMs and auxiliary limbs. Afterward, kinematics and Jacobian matrix of a typical legged lander are developed, the workspace is analyzed, singularity configurations are analyzed based on the wrench graph method, and finally, the orientation capability is analyzed. The pitch and yaw angle reach 17.5°, considering the permissible range of joints. The proposed types of LOCs offer potential candidate for the “Returning” mission of the exploration mission.     

基于逻辑回归的APU性能状况评估

针对飞机辅助动力装置(Auxiliary Power Unit,简称APU)性能评估研究中单参数评估不全面的问题,提出了基于逻辑回归模型的方法,融合多参数来对飞机APU性能状况进行全面评估.首先,对APU参数数据来源进行分析,选择基于实时报文数据的APU性能评估方法,并结合APU的工作原理以及实际工作情况,挑选出可以反...     

DFCluster: An efficient algorithm to mine maximal differential biclusters for single pilot operations task synthesis safety analysis

With the continuous advancement of the avionics system, crew members are correspondingly reduced, and Single Pilot Operations(SPO) has attracted widespread attention from scholars.To meet the flight requirements in SPO mode, it is necessary to further strengthen air-ground coordination system integration, but at the same time, there will be some safety issues caused by resource integration, function fusion, and task synthesis. Aimed at the safety problems caused by task synthesis, an efficient d...     

基于虚拟机QEMU的嵌入式全系统仿真测试环境的研究与实现

为了能节省测试设备的成本和提高嵌入式软件系统测试的效率,本论文设计了一个基于QEMU虚拟机技术的通用嵌入式全系统仿真测试环境,来实现嵌入式软件的系统测试。本文将着重介绍本测试系统中被测设备硬件仿真部分的设计与实现,以及针对该部分进行测试的测试结果。     

基于高速纹影/阴影成像的流场测速技术研究进展

本文对近年基于纹影/阴影成像的二维和三维速度场测量方法进行了综述，主要内容包括纹影成像的基本原理、硬件设备和测速算法的研究进展。在二维测速方面，介绍了纹影/阴影PIV算法、光流算法及改进算法的原理、适用场景以及优缺点。纹影特性改进光流测速算法可以实现高精度、高空间分辨率的速度场计算，适用范围相对较广。在三维粒子追踪测速方面，主要介绍了层析阴影成像、双视角平行光段阴影成像、双视角汇聚光段阴影成像三种系统的光路设置，并对各自采用的粒子重构和追踪算法进行了比较。双视角阴影成像系统的光路布置更为简洁，降低了对硬件设备的要求，在高速测量中更具优势。梳理了近年来三维粒子追踪测速算法的发展脉络，重点介绍了“先追踪–后重构”和“时间–空间耦合”的双视角三维粒子追踪测速算法。时间–空间耦合的三维粒子追踪测速算法充分利用了时间和空间信息，将时序信息引入立体匹配过程中，显著提升了双视角阴影成像系统在粒子图像密度较高时的重构正确率和追踪准确率，其整体性能优于多种人工智能算法。测速算法在上述方面取得的研究进展，结合短曝光、高帧频的图像采集优势，使得纹影/阴影成像成为一种新型的高帧频、高精度的速度测量技术，在复杂湍流及高瞬态流场实验研究中具有广泛的应用前景。     

Reinforcement learning based dynamic distributed routing scheme for mega LEO satellite networks

Recently, mega Low Earth Orbit (LEO) Satellite Network (LSN) systems have gained more and more attention due to low latency, broadband communications and global coverage for ground users. One of the primary challenges for LSN systems with inter-satellite links is the routing strategy calculation and maintenance, due to LSN constellation scale and dynamic network topology feature. In order to seek an efficient routing strategy, a Q-learning-based dynamic distributed Routing scheme for LSNs (QRLSN) is proposed in this paper. To achieve low end-to-end delay and low network traffic overhead load in LSNs, QRLSN adopts a multi-objective optimization method to find the optimal next hop for forwarding data packets. Experimental results demonstrate that the proposed scheme can effectively discover the initial routing strategy and provide long-term Quality of Service (QoS) optimization during the routing maintenance process. In addition, comparison results demonstrate that QRLSN is superior to the virtual-topology-based shortest path routing algorithm.