一种结构噪声抑制的非线性滤波器算法及模型

针对线性滤波器难以滤除超声回波信号中结构噪声的问题,从分析超声回波结构噪声和缺陷信号的产生、传播衰减及频率响应模型出发,推导了一种基于神经网络NARX(Nonlinear Auto-Regressive Exogenous Input)结构的非线性滤波器模型及其改进RTRL(Real Time Recurrent Learning)算法,该算法模型动态地建立了超声回波与缺陷信号之间的数学映射关系,利用该映射关系实现对缺陷信号中结构噪声抑制。仿真实验证实了该算法模型建立的正确性和有效性。     

基于虚拟样机技术的航天器配重设计与实现

将航天器的质量、质心和转动惯量调配到一定目标范围内,可以降低航天器姿态控制的难度,提高航天器姿态控制的准确性,利用虚拟样机技术和数值优化技术的相关软件设计了航天器配重仿真优化方法,能够对航天器进行构件设计,仿真装配、动力学分析和可变装配参数优化,能够支持构件可变质量优化和装配位置优化,最后通过空间机器人地面气浮仿真平台的配重实例验证了配重方法的正确性,结果表明该方法可以有效地对航天器进行配重优化。

Abstract：

The mass, centroid and moment of inertia of spacecraft, regulated at a certain target range, can reduce the difficulty and improve accuracy in spacecraft stance control. The relative software of virtual prototyping and numerical optimization techniques were used for designing the methods of simulation and optimization in spacecraft balance. The methods could make spacecraft devised in components, simulation assembly, dynamic analysis and parameter optimization of variable assembly and support the components of variable quality optimization and assembly location optimization. The accuracy of balancing platform was validated by the balancing example of space robot ground floatation simulation platform. The result shows that spacecraft can be effectively balanced and optimized by this method.