基于模糊模型的混沌时间序列预测

对于复杂、病态、非线性动态系统，基于模糊集合的模糊模型，利用模糊推理规则描述动态系统的特性，是一种有效方法.讨论了利用模糊建模方法实现非线性系统的建模和预测.首先，利用在线模糊竞争学习方法划分输入变量的模糊输入空间，然后利用卡尔曼滤波算法估计模糊模型的参数.采用该方法对Mackey Glass混沌时间序列进行预测试验，结果表明利用本方法可以在线或者离线能对Mackey Glass混沌时间序列进行准确预测，证明了本方法的有效性. 关键词： 模糊竞争学习 混沌时间序列 卡尔曼滤波     

Cooperatively surrounding control for multiple Euler–Lagrange systems subjected to uncertain dynamics and input constraints

In this paper, we investigate cooperatively surrounding control(CSC) of multi-agent systems modeled by Euler–Lagrange(EL) equations under a directed graph. With the consideration of the uncertain dynamics in an EL system, a backstepping CSC algorithm combined with neural-networks is proposed first such that the agents can move cooperatively to surround the stationary target. Then, a command filtered backstepping CSC algorithm is further proposed to deal with the constraints on control input and the absence of neighbors' velocity information. Numerical examples of eight satellites surrounding one space target illustrate the effectiveness of the theoretical results.     

深空天文测速自主导航速度矢量合成误差传递分析

针对深空天文测速自主导航系统速度矢量合成过程中的误差传递问题,推导了视向速度测量误差与定速误差统计特性之间的映射关系,获得了当测量误差满足零均值高斯分布时的定速误差概率密度函数,给出了在特定条件下定速误差均值与方差的解析表达式。理论与仿真分析均表明,当三个恒星视线方向两两正交时,测量误差对定速误差的影响达到最小,仿真结果还给出了目标恒星夹角对误差传递的影响,这些研究结果可为测速导航系统的目标选取提供理论依据。     

Cooperative impulsive formation control for networked uncertain Euler–Lagrange systems with communication delays

This paper investigates the cooperative formation problem via impulsive control for a class of networked Euler–Lagrange systems. To reduce the energy consumption and communication frequency, the impulsive control method and cooperative formation control approach are combined. With the consideration of system uncertainties and communication delays among agents, neural networks-based adaptive technique is used for the controller design. Firstly, under the constraint that each agent interacts with its neighbors only at some sampling moments, an adaptive neural-networks impulsive formation control algorithm is proposed for the networked uncertain Euler–Lagrange systems without communication delays. Using Lyapunov stability theory and Laplacian potential function in the graph theory, we conclude that the formation can be achieved by properly choosing the constant control gains. Further, when considering communication delays,a modified impulsive formation control algorithm is proposed, in which the extended Halanay differential inequality is used to analyze the stability of the impulsive delayed dynamical systems. Finally, numerical examples and performance comparisons with continuous algorithm are provided to illustrate the effectiveness of the proposed methods.