随机集的概率假设密度粒子滤波

多目标跟踪问题中,当目标数已知时,可以用概率数据互联(PDA)或联合概率数据互联(JPDA)算法。而当目标数未知或随时间变化时,需要对不同目标数的跟踪进行比较。可以把目标集看作随机集进行讨论,目标数N是随机变量。随机集的跟踪通过有限集统计(FISST)理论来完成。文中讨论了用粒子滤波实现跟踪随机集的方法。实验表明,在杂波环境下,粒子滤波可以稳健跟踪目标状态和目标数。     

Solar Sources of Heliospheric Energetic Electron Events—Shocks or Flares?

Electrons with near-relativistic (E≳30 keV, NrR) and relativistic (E≳0.3 MeV) energies are often observed as discrete events in the inner heliosphere following solar transient activity. Several acceleration mechanisms have been proposed for the production of those electrons. One candidate is acceleration at MHD shocks driven by coronal mass ejections (CMEs) with speeds ≳1000 km s⁻¹. Many NrR electron events are temporally associated only with flares while others are associated with flares as well as with CMEs or with radio type II shock waves. Since CME onsets and associated flares are roughly simultaneous, distinguishing the sources of electron events is a serious challenge. On a phenomenological basis two classes of solar electron events were known several decades ago, but recent observations have presented a more complex picture. We review early and recent observational results to deduce different electron event classes and their viable acceleration mechanisms, defined broadly as shocks versus flares. The NrR and relativistic electrons are treated separately. Topics covered are: solar electron injection delays from flare impulsive phases; comparisons of electron intensities and spectra with flares, CMEs and accompanying solar energetic proton (SEP) events; multiple spacecraft observations; two-phase electron events; coronal flares; shock-associated (SA) events; electron spectral invariance; and solar electron intensity size distributions. This evidence suggests that CME-driven shocks are statistically the dominant acceleration mechanism of relativistic events, but most NrR electron events result from flares. Determining the solar origin of a given NrR or relativistic electron event remains a difficult proposition, and suggestions for future work are given.     

在非结构动网格上追踪多介质界面的ALE方法

在处理带有介质分界面的多介质流动问题时,把介质边界当作一种特殊的边界,并且在该边界上的点都有着双重的定义,分别对应边界两侧的不同介质状态,通过利用HLLC方法求解任意拉格朗日-欧拉方程(ALE),以及运用界面跟踪和动网格方法分别对一维和二维多介质流进行了数值模拟,并得到了满意的结果.     

Fe/O Ratios in Interplanetary Shock Accelerated Particles

It is widely accepted that diffusive shock acceleration is an important process in the heliosphere, in particular in producing the energetic particles associated with interplanetary shocks driven by coronal mass ejections. In its simplest formulation shock acceleration is expected to accelerate ions with higher mass to charge ratios less efficiently than those with lower mass to charge. Thus it is anticipated that the Fe/O ratio in shock-accelerated ion populations will decrease with increasing energy above some energy. We examine the circumstances of five interplanetary shocks that have been reported to have associated populations in which Fe/O increases with increasing energy. In each event, the situation is complex, with particle contributions from other sources in addition to the shock. Furthermore, we show that the Fe/O ratio in shock-accelerated ions can decrease even when the shock is traveling through an Fe-rich ambient ion population. Thus, although shock acceleration of an Fe-rich suprathermal population has been proposed to explain large Fe-rich solar particle events, we find no support for this proposal in these observations.     

基于最优误差动力学的时间角度控制制导律

针对带有攻击时间和终端攻击角度约束的导弹制导问题,设计了一种基于最优误差动力学的时间角度控制制导律,并给出了明确的性能指标。推导广义最优角度控制制导律作用下的剩余飞行时间估算表达式,在广义最优角度控制制导律的基础上增加攻击时间误差反馈项,将攻击时间误差看做跟踪误差,设计的制导律使跟踪误差以最优模式在有限时间内收敛到零,最终实现攻击时间和终端攻击角度的共同控制。对不同参数下的情况进行仿真,验证了所提制导律的有效性。     

基于改进RRT算法的无人机航路规划与跟踪方法研究

为了使航路规划算法在三维动态环境下能够快速规划出较优可行航路,基于快速扩展随机树算法(RRT),对规划航路点进行了无人机飞行动力学约束,并且设计了局部航路动态优化策略。针对传统的航路跟踪控制律跟踪较为曲折的航线时跟踪误差较大的问题,通过将规划算法得出的姿态指令引入姿态控制回路的方式,提高了航路跟踪控制算法的快速性与准确性。在此基础上,搭建了无人机验证平台,利用该验证平台完成了无人机自主避障飞行试验,对算法的有效性进行了验证,并对算法性能进行了评估。     

基于视觉注意机制的认知雷达数据关联算法

针对传统关联波门设计方法在应用于机动目标跟踪时容易引起失跟、以及概率数据关联算法不适于多交叉目标跟踪的问题,提出了一种基于人类视觉选择性注意机制和知觉客体的"特征整合"理论的认知雷达数据关联算法。算法以综合交互式多模型概率数据关联算法为基础,采取假设目标最大机动水平已知的"当前"统计模型和匀速运动模型作为模型集,通过实时交互使关联波门能够随目标机动动态调整,较好地兼顾了雷达计算耗时和跟踪成功率。在利用目标位置特征的基础上,进一步提取、整合目标运动特征,对关联波门交叉区域公共量测进行分类,使多交叉目标跟踪问题转化为多个单目标跟踪问题,优化了传统概率数据关联算法。仿真结果表明：与传统关联波门设计方法相比,算法跟踪失败率和计算耗时明显降低;而且在计算资源增加不大的情况下,杂波环境适应性也得到了显著增强。     

动态时频谱分析、探测和跟踪的随机有限集法

动态出现和消失多分量信号的时频分析问题一直是非平稳信号处理的难点之一。为此,提出了一种分析、探测和跟踪多分量信号的随机有限集法。该算法利用时频变换,如短时傅里叶变换或自适应谱估计法,以及多项式预测模型,将多分量信号的时频分析问题归纳成可利用随机有限集进行多目标追踪的问题。分析表明：借助于提出的初始权重赋值算法,以及谱分量幅度和频率的联合似然函数,就可利用高斯混合概率假设密度滤波器来实现对动态时频谱的分析、探测和跟踪。在仿真实验中,所提算法有效提升了动态时频谱的跟踪精度,其对微弱时频谱分量的探测能力,以及对载频差异的分析能力均优于文献报道的算法。     

光线追踪求交算法的研究及其实现

本文对光线追踪求交算法进行了深入讨论。文中结合形体设计实例说明了此算法的正确性和实用性。     

惯导平台式重力测量技术实现及应用

详细阐述了解析方位双轴惯导平台式重力测量的工作原理,并通过基于方位余弦积的惯导力学编排解决了极区工作问题。在常规平台式重力测量数据处理方法的基础上,针对高动态环境提出了基于Kalman滤波的运动扰动修正方法。在此基础上研制了ZL11-1A型国产惯导平台式海洋重力仪。通过与国外主流重力仪产品海上同船作业比对,结果表明,该型重力仪在高海况下测量精度依然优于1mGal,满足重力测量作业要求。