多机械臂的分布式自适应迭代学习控制

针对拓扑结构为无向连通的多机械臂系统,提出了一种自适应与迭代学习相结合的分布式控制协议来实现整个系统对给定期望参考轨迹的一致性跟踪.通过引入一个适当的自适应迭代学习参数,所提自适应迭代学习控制协议能够克服机械臂系统中的干扰和模型不确定性,并且每个机械臂的自适应迭代学习控制(AILC)律仅需要利用其与邻居机械臂的相对交互信息.进一步,在只有一部分机械臂具有期望参考轨迹信息的前提下,该控制协议可以实现整个系统对期望参考轨迹的跟踪,同时能够保证轨迹跟踪误差与控制输入的有界性.此外,利用李亚普诺夫分析方法证实了所得结论的正确性,并且通过一个实例验证了所提自适应迭代学习控制协议的有效性.      

基于UKF的FADS/INS融合大气数据估计

针对飞行器在高速飞行时受气流干扰、惯性数据易发散等问题,从传感器数据融合角度出发,提出了通过无迹卡尔曼滤波(UKF)融合嵌入式大气数据观测系统(FADS)和惯性导航系统(INS)估计飞行器实时大气数据的算法。算法使用高维度非线性方程对惯性系统和大气系统间的关系建模,结合FADS与INS的数据,计算飞行器速度和高度,进而估算出攻角、侧滑角等参数。实验结果显示,与INS直接解算、扩展卡尔曼滤波(EKF)融合等原有估计方法相比,文章所述的算法在估计精度和系统稳定性方面均有所提高。     

基于模型降阶的随机最优输出反馈阵设计

本文研究了线性离散随机系统的模型降阶和最优输出反馈矩阵的设计问题。本文提出的设计方法不需要对原系统进行Anderson-Moore的迭代计算,直接利用求得的降阶模型解出近似的最优输出反馈阵。文中举例说明了设计方法的应用并进行了计算机仿真,仿真结果是令人满意的。     

Shock protection of penetrator-based instrumentation via a sublimation approach

It is often necessary for space-borne instrumentation to cope with substantial levels of shock acceleration both in the initial launch phase, as well as during entry, descent and landing in the case of planetary exploration. Current plans for a new generation of penetrator-based space missions will subject the associated on-board instrumentation to far greater levels of shock, and ways must therefore be found to either ruggedize or else protect any sensitive components during the impact phase. In this paper, we present an innovative method of shock protection that is suited for use in a number of planetary environments, based upon the temporary encapsulation of said components within a waxy solid which may then be sublimated to return the instrument back to its normal operation. We have tested this method experimentally using micromachined silicon suspensions under applied shock loads of up to 15,000g, and found that these were able to survive without incurring damage. Furthermore, quality factor measurements undertaken on these suspensions indicate that their mechanical performance remains unaffected by the encapsulation and subsequent sublimation process.     

分布式卫星系统仿真平台研究现状与技术分析

支持分布式卫星系统(DSS)开发和实验的仿真平台是DSS技术发展的关键技术之一,本文对DSS仿真平台现状和关键技术进行了分析。首先对比分析了现有3种典型的仿真平台,认为为追求尽可能高的逼真度,实时分布式仿真成为必然。继而从分布式计算角度分析了实时分布式仿真面临的分布性、实时性、仿真部件互操作性和可重用性的问题,而中间件技术可以很好地解决这些问题。最后讨论了中间件实现模式,并分析了中间件在DSS仿真中3种可能的应用方式。     

基于多滑模调节器切换的机动飞行边界保护控制

针对飞行器机动过程中关键飞行参数容易超出其边界的问题,研究了基于多滑模调节器的边界保护控制器。利用滑模方法设计多个边界调节器并采用最大/最小逻辑在各调节器之间进行切换。首先对边界约束集的正不变性以及系统最终的收敛性进行了严格的证明,然后通过将系统化为可控标准型,提出了确定系统最终收敛点的直观方法;其次通过引入分段线性滑模和分段二次Lyapunov函数对闭环系统的稳定性进行了分析;然后在此基础上给出了机动边界保护系统控制器的设计步骤;最后通过仿真表明,所设计的控制器能够保证在机动过程中关键飞行参数不越界的同时对输入指令进行很好的跟踪。     

Earth observation from space – The issue of environmental sustainability

Remote sensing scientists work under assumptions that should not be taken for granted and should, therefore, be challenged. These assumptions include the following:1. Space, especially Low Earth Orbit (LEO), will always be available to governmental and commercial space entities that launch Earth remote sensing missions.2. Space launches are benign with respect to environmental impacts.3. Minimization of Type 1 error, which provides increased confidence in the experimental outcome, is the best way to assess the significance of environmental change.4. Large-area remote sensing investigations, i.e. national, continental, global studies, are best done from space.5. National space missions should trump international, cooperative space missions to ensure national control and distribution of the data products.At best, all of these points are arguable, and in some cases, they're wrong. Development of observational space systems that are compatible with sustainability principles should be a primary concern when Earth remote sensing space systems are envisioned, designed, and launched. The discussion is based on the hypothesis that reducing the environmental impacts of the data acquisition step, which is at the very beginning of the information stream leading to decision and action, will enhance coherence in the information stream and strengthen the capacity of measurement processes to meet their stated functional goal, i.e. sustainable management of Earth resources. We suggest that unconventional points of view should be adopted and when appropriate, remedial measures considered that could help to reduce the environmental footprint of space remote sensing and of Earth observation and monitoring systems in general. This article discusses these five assumptions in the context of sustainable management of Earth's resources. Taking each assumption in turn, we find the following:(1) Space debris may limit access to Low Earth Orbit over the next decades.(2) Relatively speaking, given that they're rare event, space launches may be benign, but study is merited on upper stratospheric and exospheric layers given the chemical activity associated with rocket combustion by-products.(3) Minimization of Type II error should be considered in situations where minimization of Type I error greatly hampers or precludes our ability to correct the environmental condition being studied.(4) In certain situations, airborne collects may be less expensive and more environmentally benign, and comparative studies should be done to determine which path is wisest.(5) International cooperation and data sharing will reduce instrument and launch costs and mission redundancy. Given fiscal concerns of most of the major space agencies – e.g. NASA, ESA, CNES – it seems prudent to combine resources.     

个性化自适应课件生成器的研究

分析基于Internet的远程教育学习站点存在两个明显的缺点,提出个性化自适应课件生成器的概念,并针对其关键技术问题进行研究与讨论。     

基于全解耦奇偶方程的动态系统执行器故障检测与识别

给出了基于全解耦奇偶方程的动态系统执行器故障的检测与识别方法。讨论了全解耦奇偶向量的产生方法,给出了全解耦奇偶向量的存在条件,并结合卡尔曼滤波方法得到了故障模型参数的估计方法。最后给出了仿真实例。结果表明,全解耦奇偶方程方法能估计出动态系统执行器故障模型。