Semi-supervised gear fault diagnosis using raw vibration signal based on deep learning

In aerospace industry, gears are the most common parts of a mechanical transmission system. Gear pitting faults could cause the transmission system to crash and give rise to safety disaster. It is always a challenging problem to diagnose the gear pitting condition directly through the raw signal of vibration. In this paper, a novel method named augmented deep sparse autoencoder (ADSAE) is proposed. The method can be used to diagnose the gear pitting fault with relatively few raw vibration signal data. This method is mainly based on the theory of pitting fault diagnosis and creatively combines with both data augmentation ideology and the deep sparse autoencoder algorithm for the fault diagnosis of gear wear. The effectiveness of the proposed method is validated by experiments of six types of gear pitting conditions. The results show that the ADSAE method can effectively increase the network generalization ability and robustness with very high accuracy. This method can effectively diagnose different gear pitting conditions and show the obvious trend according to the severity of gear wear faults. The results obtained by the ADSAE method proposed in this paper are compared with those obtained by other common deep learning methods. This paper provides an important insight into the field of gear fault diagnosis based on deep learning and has a potential practical application value.     

基于DOA聚类算法的ARM抗有源诱偏技术研究

有源诱偏技术是雷达对抗反辐射导弹的一种重要手段,可以大大降低反辐射导弹的作战效能。为提高对敌目标的精确打击能力,在分析闪烁诱饵诱偏原理的基础上,针对有源诱偏干扰下被动雷达测角精度与稳定度不高的问题,通过对有源诱偏信号的时域特征进行分析,提出一种基于脉冲前沿检测的DOA聚类分选算法,找出前沿超前的辐射源信号,实现了高性能抗有源诱偏干扰和高精度抗干扰测向。仿真结果表明,该技术可以有效对抗四点源有源诱偏干扰。     

联邦学习安全与隐私保护综述

联邦学习是一种新型的分布式学习框架，它允许在多个参与者之间共享训练数据而不会泄露其数据隐私。但是这种新颖的学习机制仍然可能受到来自各种攻击者的前所未有的安全和隐私威胁。本文主要探讨联邦学习在安全和隐私方面面临的挑战。首先，本文介绍了联邦学习的基本概念和威胁模型，有助于理解其面临的攻击。其次，本文总结了由内部恶意实体发起的3种攻击类型，同时分析了联邦学习体系结构的安全漏洞和隐私漏洞。然后从差分隐私、同态密码系统和安全多方聚合等方面研究了目前最先进的防御方案。最后通过对这些解决方案的总结和比较，进一步讨论了该领域未来的发展方向。     

在片校准标准件基底边界条件影响

通过分析国内外校准标准件物理边界条件，研究不同校准基底的物理边界条件对共面波导（CPW）传输线特性和校准准确度的影响。研制了W波段砷化镓衬底的多线TRL在片校准标准件，将在片校准标准件分别放置在金属卡盘、玻璃片、吸波材料三种校准基底上。用三种不同校准基底的校准标准件对试验装置校准后，分别提取了CPW传输线的衰减常数、相对介电常数和特征阻抗，并对商用校准标准件104-783A的短路标准和传输线标准进行了测试，验证了结果的合理性。     

Fully distributed time-varying formation tracking control for multiple quadrotor vehicles via finite-time convergent extended state observer

This paper investigates a time-varying anti-disturbance formation problem for a group of quadrotor aircrafts with time-varying uncertainties and a directed interaction topology. A novel Finite-Time Convergent Extended State Observer (FTCESO) based fully-distributed formation control scheme is proposed to enhance the disturbance rejection and the formation tracking performances for networked quadrotors. By adopting the hierarchical control strategy, the multi-quadrotor system is separated into two subsystems: the outer-loop cooperative subsystem and the inner-loop attitude subsystem. In the outer-loop subsystem, with the estimation of disturbing forces and uncertain dynamics from FTCESOs, an adaptive consensus theory based cooperative controller is exploited to ensure the multiple quadrotors form and maintain a time-varying pattern relying only on the positions of the neighboring aircrafts. In the inner-loop subsystem, the desired attitude generated by the cooperative control law is stably tracked under a FTCESO-based attitude controller in a finite time. Based on a detailed algorithm to specify the cooperative control protocol, the feasibility condition to achieve the time-varying anti-disturbance formation tracking is derived and the rigorous analysis of the whole closed-loop multi-quadrotor system is given. Some numerical examples are conducted to intuitively demonstrate the effectiveness and the improvements of the proposed control framework.     

Optimization of bits allocation and path planning with trajectory constraint in UAV-enabled mobile edge computing system

In this paper, an Unmanned Aerial Vehicle (UAV) enabled Mobile Edge Computing (MEC) system is studied, in which UAV acts as server to offer computing offloading service to the Mobile Users (MUs) with limited computing capability and energy budget. We aim to minimize the total energy consumption of MUs by jointly optimizing the bit allocation for uplink, computing at the UAV and downlink, along with the UAV trajectory in a unified framework. To this end, a trajectory constraint model is employed to avoid sudden changes of velocity and acceleration during flying. Due to high-order information in use, we lead to a more reasonable nonconvex optimization problem than prior arts. An Alternating Direction Method of Multipliers (ADMM) method is introduced to solve the optimization problem, which is decomposed into a set of easy sub-problems, to meet the requirement on the efficiency in edge computing. Numerical results demonstrate that our approach leads a smoother UAV trajectory, significantly save the energy consumption for UAV during flying.     

基于梯度提升决策树的量子科学实验卫星光学实验预测

量子科学实验卫星在轨运行期间完成4种光学实验，地面监测人员通过遥测参数阈值判断卫星是否进行光学实验、实验类型及实验结果.这种方法需要预先设定大量阈值，并且这些阈值需要根据在轨卫星重新设定，可扩展性较差.针对以上问题，提出一种基于机器学习的光学实验判别方法，将量子科学实验卫星的光学实验监测任务抽象为机器学习中的多元分类问题，构建分类模型，利用量子科学实验卫星的真实历史遥测数据对模型进行训练，并通过真实实验计划对训练得到的模型进行验证.实验结果表明，本文提出的方法在没有专家先验知识的前提下，判别准确率达到99%，可用于量子科学实验卫星光学实验的实时监测任务.提出的基于机器学习的判别方法具有较强的可扩展性，可应用于卫星在轨运行的其他监测任务.      

弱引力下小行星动量驱动机器人的可控跳跃行为

以小行星表面着陆探测为背景,提出一种动量驱动机器人(MoRo)以满足弱引力复杂环境下的探测需求。该机器人利用弱引力环境下的摩擦和碰撞特性,通过主动辨识环境参数,规划和控制动量轮以产生期望的驱动力矩,完成可控性跳跃及腾空后的稳定拍照等任务。首先,基于MoRo的动量轮刹车机构特性,分析了MoRo在弱引力环境下的跳跃机理并对其跳跃方式进行了规划;接着考虑动量轮驱动机构三闭环伺服系统的非线性特性,基于Herze碰撞模型和Karnopp摩擦模型建立了MoRo在小行星表面的跳跃行为动力学模型;其次,使用机器学习算法建立环境参数和MoRo运动的函数关系,并基于环境参数规划动量轮转速实现跳跃距离和腾空高度的可控。最后,通过数值仿真校验了MoRo跳跃规划方法和控制方法的可行性。     

单向电永磁作动器结构优化与动态特性分析

针对电磁作动器的使用要求，设计出一种具有能耗低、吸力大及响应快等特点的新型单向电永磁作动器。结合有限元仿真软件，利用有限元法研究作动器的结构参数如铁芯齿形结构、平面结构、无永磁结构、铁芯的倾斜角度及衔铁位移（气隙厚度）等因素对作动器电磁吸力和响应速度的影响。结果表明，电磁吸力随着作动衔铁位移的增加而大幅度减小；倾角结构的铁芯设计能使作动器性能达到最佳，且铁芯的倾斜角度为50°时，电磁吸力有最大值3071N。