超分辨测高的试验研究

多径信号的存在,使得米波雷达在低仰角的测高成为难题。随着空间谱估计技术的发展,人们尝试采用阵列超分辨方法实现米波雷达的仰角维方向估计。简要介绍了常规测高法的局限性和角度超分辨实现测高的基本原理,以及在米波雷达中开展的测高试验情况,给出了若干淡水反射面条件下的外场试验数据。对多批目标的试验统计结果表明:在快拍数只有一个、单个阵元平均信噪比达到8dB以上时,采用空间平滑+最大似然超分辨处理法即可有效分离相干信号源、减小米波雷达在低仰角的测高误差。     

超分辨算法在米波雷达测高中的应用

简要介绍了超分辨算法实现测高的基本原理,以及采用超分辨算法开展的米波雷达测高试验情况,给出了若干淡水反射面条件下的外场试验数据。试验统计结果表明：在满足一定信噪比的条件下,采用空间平滑＋最大似然超分辨处理法可有效提高米波雷达在低仰角的测高精度。     

广义MUSIC算法在米波雷达测高中的应用及其改进

米波雷达波束较宽,多径信号的存在将会给雷达测高带来严重影响,并且多径信号与直达波信号是相干的,这就造成米波雷达测高比较困难。由于广义MUSIC算法为能直接处理相干信号的阵列超分辨算法,因此可以将该算法应用于米波雷达测高当中。该文在广义MUSIC算法的基础上,结合雷达测高的一些特点,提出一种新的算法。新算法不但降低了运算量,而且还具有良好的性能。计算机仿真结果证实了新算法的优越性。     

米波雷达超分辨的实验研究

简要地介绍了角度超分辨的基本原理以及米波雷达超分辨体制试验站的组成和主要技术指标,提出了解决通道修正、运动补偿等关键技术的方法。最后给出了米波雷达角度超分辨的外场实验结果。     

一种机载/弹载阵列雷达前视超分辨成像算法

针对机载/弹载雷达前视成像过程中方位分辨率急剧下降的问题,提出了一种基于超分辨技术的阵列雷达前视成像算法。该算法通过对接收的各通道回波数据进行样本协方差估计,并对该协方差进行修正及空间平滑等处理,随后采用改进的多重信号分类(MUSIC)算法估计空间谱,再根据系统参数将谱曲线进行累积,最终获得雷达前视图像。仿真及实测数据处理结果表明,该算法能有效提高雷达前视方向成像的方位分辨率,实现机载/弹载雷达前视超分辨成像。     

用低维子空间法分辨空间相干源的统计分析

八十年代以来,Schmidt提出的基于特征矢量空间的多信号分类(MUSIC)法受到广泛关洼。Evans,Shan等采用取较短的子阵列,并作空间平滑处理,将MUSIC法推广到相干源的场合。文献[7]、[10]、[14]从数据矩阵出发,提出了构造低维子空间以分辨相干源的一般方法。在诸多方法中,它们的分辨性能如何,以及怎样进一步提高分辨率需要深入研究,本文通过统计分析研究了上述问题。     

空间平滑差分方法

针对相干信号源超分辨空间谱估计中存在的问题，本文提出了空间平滑差分方法。利用常规超分辨方法估计不相关源和相关源的空间谱，利用空间平滑差分方法估计相关源和相干源的空间谱，这样可以重复利用接收阵列输出数据，兼顾不相关信号源和相干信号源的空间谱估计，因而可以估计更多的信号源。     

基于实数遗传算法的米波雷达方位超分辨研究

超分辨谱估计技术可以在不增大天线口径的情况下提高米波雷达的方位角分辨力.利用实数遗传算法实现了波达方向(DOA)估计的参数模型拟合类超分辨方法,并利用计算机仿真和米波雷达的实测数据进行了算法性能分析,分析结果表明利用实数遗传算法的全局寻优能力可以改进米波雷达方位超分辨方法的性能.     

基于射线追踪的米波雷达低角测高

由于实际地形条件很复杂,难以准确建立雷达信号多径传播模型,从而影响了米波雷达低角测高的性能。文中提出采用射线追踪法来改进雷达反射信号多径模型,使其更接近于电波真实传播路径,以期达到提高超分辨测高算法地形适应性的目的。仿真和不同地形条件下的试验结果表明：基于改进后的多径反射模型的超分辨算法在解决地形适应性问题上卓有成效。     

改进SSMUSIC超分辨多径时延估计算法

传统的MUSIC超分辨时延估计技术是直接基于测量数据,其性能往往只对宽带且频谱近似平坦的信号较优,而对窄带信号估计性能较差。针对上述问题,本文通过利用谐波频率估计模型和DOA (Direction of Arrival)估计模型之间的等价性,将时延估计问题转化为谐波频率估计问题,提出了一种改进SSMUSIC(Signal Subspace Scaled Multiple Signal Classification)超分辨多径时延估计算法。改进后的算法采用平滑的思想和SSMUSIC算法的思想构造协方差矩阵和MUSIC谱,实现了对多径时延的超分辨估计。仿真表明,该算法能够实现对窄带信号多径时延超分辨估计且具有DP(Direct-Path)不模糊和谱峰陡峭的特点,估计性能优于传统的超分辨算法。      

一种基于双空间密度的多目标进化算法

目前,大多数多目标进化算法的多样性保持机制都只强调目标空间的多样性而忽视决策空间的多样性.这种设置可能导致种群在目标空间拥有良好多样性的同时却在决策空间拥有较差的多样性.为了解决上述问题,本文提出了一种基于双空间密度的多目标进化算法.为了反映个体在决策空间和目标空间的多样性,本文定义了双空间密度指标.基于双空间密度的配对选择操作可以平衡算法的收敛性与多样性;基于双空间密度的个体选择操作可以同时使决策空间和目标空间得到充分的搜索.实验结果表明,本文算法的求解质量明显优于对比算法.     

核不相关空间算法及其在人脸识别中的应用

不相关空间算法是求解不相关鉴别矢量集的快速算法,但是将其应用在人脸识别中将遇到小样本问题,并且算法只是一种线性的特征提取方法。该文提出一种核不相关空间算法,该方法的关键是高维特征空间中不相关空间的计算,对此提出一种简单的计算方法,即根据eigenface中将高阶矩阵计算转化成低阶矩阵计算的思想,将高维特征空间中不相关空间的计算仍归结为标准的特征值分解问题。所提出的算法能够有效地解决小样本问题。在ORL人脸库上的实验结果验证了所提出的算法的可行性和有效性。     

一种基于状态空间的启发式搜索算法及其实现

深度优先和广度优先搜索算法由于需遍历所有状态空间才能求出最佳解,使其在状态空间较大时效率极低,此时必需采用启发式算法实现快速求解。阐述启发式搜索算法在状态空间较大时的广泛应用,深入分析一种启发式算法-A-Star算法实现快速求解的原理,并详细介绍了其实现步骤及过程。最后,得出结论:基于合理估价函数的A-Star算法能极大提高求解效率。     

基于语义分析的互联网人物信息提取

互联网是人类网络空间行为的体现,其中隐藏了大量人物信息.由于这些信息分散在整个网络空间中,将互联网人物信息提取并进行归类具有重要的研究意义和实用价值.文中提出了一种新的互联网人物信息提取模型,实现了人物信息的自动化提取.详细分析了基于网络爬虫的网页信息采集、基于语义分析的人物特征提取、基于向量空间模型的人物聚类算法和人物信息检索等技术原理和实现方案,能够对互联网人物信息进行分析和提取.     

A 10-Bit,Low Power,Successive Approximation,Digitally Auto-Zeroed CMOS ADC Core for the NASA TRIO Smart Sensor System on a Chip

In spacecraft applications there is a great need for robust analogue to digital converters (ADC) that can withstand the harsh space environment. Commercially available ADCs cannot operate in the space environment due to radiation effects. In this paper we present an ADC that has been developed for the NASA TRIO smart sensor system on a chip (SoC), a versatile low power device specifically designed for spacecraft data acquisition and telemetry of several types of sensors such as temperature, voltage/current transducers, radFETs, etc. It is required for the ADC to operate in excess of 300 Krad total ionizing dose and to be robust to single event upsets. The successive approximation topology was chosen and it was enhanced with a special auto-zeroing technique to compensate for possible lifetime offset errors. Due to the comparator design, a rail-to-rail input capability is achieved, a feature very useful in some type of Vdd ratio metric sensors. It has 10-bit resolution for a reference in the range 0.1 to Vdd + 1 V, and for power supply in the range 2.5 to 5.5 V; the positive reference terminal V_ref+ is settable up to Vdd + 0.5 V and the negative voltage terminal is settable down to GND-0.5 V. The power dissipation is less than 2 mW at 50 Ksamlles/sec. The TRIO chip is used in several NASA spacecraft including CONTOUR, STEREO, MESSENGER, EUROPA, PLUTO, etc.George Kottaras was born in Athens, Greece in 1974. He received the Diploma degree (five years with thesis) in Electrical Engineering from Democritos University of Thrace, Greece in 1996. He is currently pursuing the Ph.D. degree on Scientific Space Instruments and spacecraft avionics at Space Research Laboratory, DUTh. He has specialized in VLSI technologies at JHU/APL for about five years.His research interests include mixed signal analog/digital design, ADCs, design for testability, testing, smart sensors and data acquisition.Nikolaos P. Paschalidis was born is Serres, Greece in 1963. He received the Diploma and Ph.D. degrees in Electrical Engineering from the Democritus University of Thrace (DUTh), Greece, in 1985 and 1992 respectively. He has been in appointment with the Johns Hopkins University, since 1989, where his research specialized in advanced microelectronics, space instrumentation, and space physics.He later joined the Space Department of JHU Applied Physics Laboratory (JHU/APL) Laurel, MD, as a postdoctoral fellow and presently he is Principal Staff. His research interests are in analog and mixed signal microelectronics, microsensors, microsystems and their applications in in-situ and remote sensing spacecraft instruments and avionics. He pioneered in the Advanced Technology Development program of NASA for smaller better faster missions by leading efforts in the circuit level of: amplifiers, comparators, voltage references, ADC and DAC, PLLs, TDCs, SEU and radiation tolerant design, physical design, design for testability, testing and space qualification; in the system on a chip level flight ready chips including: the Time of Flight chip for precise time pickoff and time digitization, Energy chip for radiation energy measurement, the TRIO smart sensor chip for spacecraft data acquisition and control etc; in the instrument and spacecraft level: application of these technologies in particle and plasma spectrometers, laser altimeters, photon/particle imagers, TOF mass spectrometers, X-ray and gamma-ray instruments, spacecraft avionics. Space missions using these technologies include: Cassini, Image, Contour, Messenger, Pluto, Mars missions, etc. Dr. Paschalidis published extensively in microelectronics, space instrumentation, and space physics. He supervises research of graduate students in ECE and Applied Physics. He supervised DUTh graduate students at JHU/APL for many years. He participates as principal investigator and co-investigator in several space programs; he participates in communities with space related activities including: the IEEE Aerospace, Nuclear Sciences, NASA VLSI, IAA, and American Geophysical Union.Emmanuel T. Sarris was bom in Athens, Greece, in 1945. He received the physics degree from the University of Athens in 1967 and the Ph.D. degree in space physics from the University of Iowa, Iowa City, in 1973.He was a Postdoctoral Fellow in the Applied Physics Laboratory, The Johns Hopkins University, Baltimore, MD, from 1974 to 1976. From 1976 to 1977, he was a Research Scientist at the Max-Planck-Institut. He has been a Professor of Electrodynamics, Department of Electrical Engineering, University of Thrace, Greece, and Director of the Laboratory of Electrodynamics and Space Research since 1977. He was the Director of the Institute of Ionospheric and Space Physics, National Observatory of Athens from 1990 to 1996. His research interests include space plasma electrodynamics, design, construction, and testing of space instrumentation, satellite communications, satellite remote sensing. He is coinvestigator in the international space missions: Ulysses, Geotail, Interball, Cluster. He is the author of 270 refereed publications and 300 presentations at international meetings. Dr. Sarris is a member of the COSPAR Council. He was elected Johns Hopkins Scholar Award in 1992 and received the Award for Academic Excellence in 1994.Nikos Stamatopoulos was born in Peloponnisos, Greece in 1969. He received the diploma degree (five years with thesis) of Electrical Engineering from Democritos University of Thrace, Greece in 1994. He is currently pursuing the Ph.D. degree on Scientific Space Instruments at Space Research Laboratory, DUTh. He has specialized in VLSI technologies with emphasis in low noise analog design at JHU/APL for about five years.His main research interests are on Analogue CMOS VLSI design for fast time acquisition.Kostas Karadamoglou was born in Macedonia, Greece, in 1970. He received the diploma degree (five years with thesis) of Electrical Engineering from Democritos University of Thrace, Greece in 1994. He is currently pursuing the Ph.D. degree on Scientific Space Instruments at Space Research Laboratory, DUTh. He has specialized in VLSI technologies with emphasis in high-speed digital design at JHU/APL for about five years.His main research interests are on the design of application specific Time to Digital Converters.Vassilis Paschalidis was born in Serres, Greece in 1964. He received the B.S. degree in Electrical Engineering from the Technological Institute of Kabala, Greece in 1988. He worked n the industry for electronic automation. He has specialized in VLSI technologies at JHU/APL for about five years with emphasis in physical design. His research interests include mixed signal analog/digital VLSI design.     

深空通信中的线性Log-MAP算法

Turbo码由于具有逼近香农信道容量限的性能,已经被CCSDS标准采纳。译码算法是Turbo码译码的关键组成部分,也是影响其性能的关键因素,而MAP类译码算法以其优越的性能成为译码算法的主体。在对MAP类译码算法的译码复杂度和性能仿真对比、分析的基础上,针对深空通信中的低信噪比、低误比特率的要求,提出一种线性Log-MAP算法方案,该方案在保证性能损失尽可能小的前提下,有效降低了计算复杂度,适用于深空通信的实际要求。     

SWEDE算法提高模式空间内DOA估计抗噪声能力

均匀测向圆阵相干信号DOA估计在模式空间内将阵列流形化为线性结构后 ,一般情况下都是先进行空间平滑去相干处理后再运用MUSIC算法进行DOA估计 ,文中运用SWEDE算法代替MUSIC算法进行了模式空间内的DOA估计 ,仿真结果证明SWEDE算法更好地适应了均匀圆阵模式空间内DOA估计数据矩阵的特点 ,使系统抗噪声能力、测向精度、分辨率均得到较大提高     

空间网络多路径路由算法

针对空间网络具有距离远、延时大、周期性间歇连接的特点,提出了一种适用于空间网络的多路径最大吞吐量的路由算法（SMMT）。该算法是对最小费用最大流算法的改进,经过多次最小费用路径查找和残留网络的构造过程,最终可以找出所有满足传输条件的转发路径。仿真实验表明,和传统的ASCOT和S-OSFP单路径路由算法相比,SMMT算法明显提高了网络的吞吐量,减少了数据包的传输时延,使得网络的资源得到充分的利用。     

主动空间映射算法设计滤波器

应用ADS和HFSS软件,分别建立了一种PCB(Printed Circuit Board)多层滤波器的粗糙模型和精确模型,采用主动空间映射算法,对其进行优化设计,制作出样品,并进行了检测。结果表明,该滤波器的中心频率为1.85 GHz,带宽为300 MHz,插入损耗小于2.3 dB,证明了主动空间映射算法的可行性。     

基于空间映射的RF MEMS 开关建模研究

空间映射的思想是通过构造粗糙模型设计变量与精确模型设计变量之间的映射关系,获得合适的替代模型（校准后的粗糙模型）,从而简化优化过程。本文针对电容耦合式RF MEMS 开关使用初始空间映射算法进行计算和优化,经过迭代,建立了精确空间和粗糙空间的映射关系,最终得到了精确空间优化设计值,利用此映射关系,可将精确空间的优化工作放到粗糙空间进行,从而大大节省时间和电脑资源。计算结果验证了该方法的可行性和高效性。