基于蒙特卡罗方法的飞机成本参数法估算研究

研发早期的成本估算是项目立项和设计决策的基础,参数法可以用最少量的信息估算成本,而蒙特卡罗方法可以进行成本数据的不确定性分析,以明确成本风险。本文研究了参数法和蒙特卡罗方法在飞机成本估算中的应用。     

基于四线性分解的双基地MIMO雷达的角度和多普勒频率联合估计

研究双基地多输入多输出(Multiple-input Multiple-output, MIMO)雷达中的角度和多普勒频率联合估计问题,提出了一种基于四线性分解(Quadrilinear Decomposition)的离开角(Direction of Departure, DOD)、波达角(Direction of Arrival, DOA)和多普勒频率的联合估计算法。通过对接收端匹配滤波器的输出进行延迟操作,得到符合四线性模型的数据,根据四线性交替最小二乘(Quadrilinear Alternating Least Squares, QALS)进行迭代,得到方向矩阵和多普勒频率矩阵的估计,进而得到角度和频率的估计。该算法无需谱峰搜索,无需知道反射系数,可实现角度和频率的自动配对,且能用于非均匀阵,该算法的角度估计性能优于多维ESPRIT方法和三线性交替最小二乘(Trilinear Alternating Least Squares, TALS)方法。论文分析了所提算法复杂度,并推导了克拉美-罗界(Cramer-Rao bound, CRB)。仿真结果验证了该算法的有效性。     

使用电场传感器阵列的水下运动目标检测与参数估计

研究了基于电场传感器阵列的水下运动目标检测与参数估计方法。在背景噪声为高斯白噪声和目标作匀速直线运动的假设下,利用电场传感器阵列获取的船舶水下准静态电场空时特征,通过广义似然比检测和随机搜索寻优算法,可实现对海水中运动船舶的目标检测和参数估计。仿真试验结果表明:在信噪比大于0 dB时能够取得较好的估计精度;在信噪比低至10 dB时仍能有效检测目标。     

某型涡轴发动机性能衰减与部件退化评估

为在某型涡轴发动机持久试验中进行性能健康检查,采用多点匹配方法建立了个体发动机性能衰减计算模型,基于非线性气路分析方法对部件性能退化进行了计算,集合已有经验确定性能偏离部件,完成了某型发动机持久试验中不同阶段的性能衰减评估.结果表明,某型发动机性能衰减及部件性能退化成阶段性分布,前期压气机和燃气涡轮性能变化最快,随后处于相对稳定阶段,动力涡轮性能在目前试验时数范围内基本无变化.该方法的验证为涡轴发动机使用过程中的性能健康与评估提供了有益的经验.      

基于瞬时相位的BPSK信噪比估计算法

通过对加性高斯白噪声条件下正弦型信号瞬时相位分布情况进行研究,提出了一种新的BPSK(Binary Phase Shift Keying)信噪比估计方法.含噪信号瞬时相位的方差与信噪比之间存在近似线性的一一对应关系,以此为基础通过将含噪信号的估计方差与理论值相比较获得信噪比估计值.该方法直接对基带过采样信号进行处理,不需要事先进行符号定时与同步.使用计算机仿真测试了算法的估计性能,并与高阶累积量方法和子空间分解方法进行了比较,结果表明该算法具有更高的估计精度.      

Simulations of absorbed dose on the phantom surface of MATROSHKA-R experiment at the ISS

The health risks associated with exposure to various components of space radiation are of great concern when planning manned long-term interplanetary missions, such as future missions to Mars. Since it is not possible to measure the radiation environment inside of human organs in deep space, simulations based on radiation transport/interaction codes coupled to phantoms of tissue equivalent materials are used. However, the calculated results depend on the models used in the codes, and it is therefore necessary to verify their validity by comparison with measured data. The goal of this paper is to compare absorbed doses obtained in the MATROSHKA-R experiment performed at the International Space Station (ISS) with simulations performed with the three-dimensional Monte Carlo Particle and Heavy-Ion Transport code System (PHITS). The absorbed dose was measured using passive detectors (packages of thermoluminescent and plastic nuclear track detectors) placed on the surface of the spherical tissue equivalent phantom MATROSHKA-R, which was exposed aboard the ISS in the Service Zvezda Module from December 2005 to September 2006. The data calculated by PHITS assuming an ISS shielding of 3 g/cm² and 5 g/cm² aluminum mass thickness were in good agreement with the measurements. Using a simplified geometrical model of the ISS, the influence of variations in altitude and wall mass thickness of the ISS on the calculated absorbed dose was estimated. The uncertainties of the calculated data are also discussed; the relative expanded uncertainty of absorbed dose in phantom was estimated to be 44% at a 95% confidence level.     

Precise line-of-sight vector estimation based on an inter-satellite radio frequency system

This paper proposes a precise line-of-sight (LOS) vector estimation using an inter-satellite radio frequency system. GNSS-like technology is inherited such that the ranging signals are locally generated inside the formation. However, the approach differs from the standard GNSS model usage in that the LOS vector to be of a unit length is fully explored as a priori constraint for the carrier phase integer ambiguity resolution. The constraint is lumped to the mapping process from the real-valued ambiguities to the integers by what is called validation or subset ambiguity bounding. These two approaches have the same rules of regarding the constraint as a gateway to accept or reject the ambiguity candidates, but differ by using “all-ambiguity-set” and “subset-ambiguity”. Both show remarkable improvement with up to 80% lower integer fixing failure rates than without treating the constraint. Validation provides a slightly better performance than the subset ambiguity bounding in terms of the integer fixing failure rates and the computational efficiency. The predefined tolerance regions that are critical for these two methods are analytically determined as function of the carrier noise. The paper also introduces a LOS dependent ambiguity dilution of precision (ADOP_LOS) measure that can serve as a metric to characterize the expectation of being able to successfully resolve the ambiguities. The region of ADOP_LOS lower than 0.21 is empirically summarized as the safe region where the integer fixing failure rates are less than 1%. A closed form of the ADOP_LOS is derived which is able to capture the impact of the various factors. Antenna baseline geometries and multiple frequencies in the form of an ultra-BOC signal structure are demonstrated as the most important influencing factors. With multiple properly arrayed antennas and using ultra-BOC structure, instantaneous ambiguity resolution can be achieved and the LOS accuracy can reach millimeter level.     

基于UKF参数估计的地月自由返回轨道设计

为快速简便地设计地月自由返回轨道,提出了一种基于UKF参数估计算法的地月自由返回轨道设计方法。该算法不仅避免了传统数值方法推导相关梯度矩阵的复杂性,而且只需基于地月系统二体模型给出猜测初值,从而显著降低了自由返回轨道设计的难度,将地月自由返回轨道对应的两点边值问题的求解转化为参数估计问题,该算法可以得到高精度模型下收敛的精确解。数值仿真结果表明:该算法结构简洁,求解效率较高,所得结果精确且具有良好的鲁棒性,可以作为地月自由返回轨道设计的一个有力工具。     

卫星自主导航系统存在常值偏差的滤波方法研究

由于敏感器常值偏差对导航精度影响较大,因此有必要对其进行标定。考虑到传统常偏扩维算法计算量大,且可能出现数值病态的问题,提出了两步UKF算法,该算法对状态和偏差实施分离估计,从而达到解耦目的。不仅能够准确标定常值偏差,而且还可以提高导航精度。近地卫星自主导航系统仿真结果验证了该算法的有效性。     

Higher order ionospheric propagation effects on GPS radio occultation signals

With the increasing number of remote sensing satellites using the GPS radio occultation technique for atmospheric sounding, the estimation of higher order ionospheric effects and their mitigation have become relevant and important. Due to long ionospheric limb paths, GPS signals are strongly affected by ionospheric refraction during radio occultation. Standard dual-frequency GPS measurements may be used to estimate the first order term of the refractive index. However, non-linear terms such as the second and third order ionospheric terms and ray path bending effects are not considered in occultation measurements so far. Analysing selected CHAMP–GPS occultation events different higher order ionospheric terms are estimated and their effects on dual-frequency range estimation and total electron content (TEC) estimation are discussed. We have found that the separation between the GPS L1 and L2 ray paths exceeds the kilometer level during occultation for a vertical TEC level of more than 160 TEC units. Corresponding errors in the GPS dual-frequency range estimation and TEC estimation are found to exceed the meter and 10 TEC units level, respectively.