Limitations Of The Tangent Plane Approximation For Wide-field Imaging Using The Mauritius Radio Telescope

The Mauritius Radio Telescope (MRT) is a Fourier synthesis, non-coplanar T array. The primary objective of the telescope is to survey the southern sky at 151.5 MHz in the declination range -70^° to -10^°. Due to non-coplanarity, wide-field imaging and deconvolution of wide field images made using MRT are challenging problems in applications of radio interferometric techniques. This paper discusses the usefulness and limitations of the tangent plane approximation in transforming the measured visibilities to wide field images and in estimating the point spread function (PSF) required for the deconvolution. This revised version was published online in July 2006 with corrections to the Cover Date.     

Uniformly Redundant Arrays

Uniformly Redundant Arrays (URA's) are two–dimensional binary arrays with constant sidelobes of their periodic autocorrelation function. They are widely agreed upon to be optimum mask patterns for coded aperture imaging, particularly in imaging systems with a cyclic coded mask. In this paper, a survey of all currently known construction methods for URA's is given and the sizes and open fractions of the arrays resulting from each construction method are pointed out. Alternatives to URA's for situations in which a URA does not exist, are discussed.     

Degradation due to Quantization Noise in Radio Astronomy Phased Arrays

In this article, a model is developed for determining the probability distribution of the output of a digital adder in case of 2-, 3- and 4-level quantization before summation. This probability distribution is then used to determine the efficiency of a system which determines the total power of the signal at the output of the digital adder.The sum-signal is also used in a VLBI correlation mode where the sum-signal is correlated with the signal from a VLBI partner. The efficiency, expressed in so-called Degradation Factors, is determined for 1- and 2-bit correlation and 2-, 3- and 4-level quantization before summation.Furthermore, the application of the calculated values (Total Power Detector Efficiencies and Degradation Factors) is illustrated.     

射频干扰检测的SumThreshold算法

射频干扰(Radio Frequency Interference, RFI)是射电目标搜寻和精确分析研究的关键影响因素,因此RFI检测是相关数据处理中的一个重要环节.已有RFI检测方法可分为成分分解法、阈值分析法、机器学习类方法.从应用广泛性和可解释性方面考虑,阈值分析法最具代表性,特别是SumThreshold是近年备受关注的一个RFI检测方法.从SumThreshold方法的原理、算法设计、优化要点、适用性等方面进行介绍和探讨,以供同行参考.     

System Optimisation Of Multi-Beam Aperture Synthesis Arrays For Survey Performance

Investigating the requirements for an aperture synthesis array that optimise the performance for surveying shows that, next to collecting area and system temperature, the field-of-view (FoV) is key parameter. However, the effective sensitivity not only depends on bandwidth and integration time but could be seriously limited by side lobe confusion and by gain errors that determine the effective dynamic range. From the basic sensitivity equation for a radiometric system we derive an optimum cost ratio between collecting area and processing electronics, where the latter should be less than a third of the total cost. For an instrument that has to cover a fraction of sky larger than its field-of-view, the FoV enters the equation for survey sensitivity and we identify the number of independent feed systems per unit collecting area as a key parameter. Then the optimum cost distribution allows the electronics to account for up to half the total cost. Further analysis of the sensitivity formula shows that there is an optimum design frequency for a survey instrument below 1 GHz. We discuss the impact of station size and array configuration on self-calibration, side lobe confusion and effective sensitivity and conclude that a minimum station size of 20 m diameter is required at 0.3 GHz as long as multi-patch self-calibration procedures need, per baseline, a signal-to-noise ratio of more than two for each ionospheric coherence patch.     

Fiber Optic Network Technology For Distributed Long Baseline Radio Telescopes

The construction costs of distributed radio telescopes are to a great extent determined by the deployment costs of the fiber optic data transport network that is needed to transport the received information to the data processor(s). As such, the baseline and data rates that are feasible for a specified amount of money are determined by the status of the technology and deployment costs of the communication network. In this paper the present day data transport status is described and, using a costing model, the most attractive data transport technologies are determined, taking the LOFAR telescope (ASTRON, 2005) as an example. In the outlook, the near-term data transport technology developments are described.     

Research on Wide-field Imaging Techniques for Low-frequency Radio Arraytwo

Wide-field imaging of low-frequency radio telescopes is subject to a number of difficult problems. One particularly pernicious problem is the non-coplanar baseline effect. It will lead to distortion of the final image when the phase of $w$-direction called $w$-term is ignored. The image degradation effects are amplified for the telescopes with a wide field of view. This paper summarizes and analyzes several $w$-term correction methods and their technical principles. Their advantages and disadvantages are analyzed after comparing their computational cost and computational complexity. We conduct simulations with two of these methods, i.e., faceting and $w$-projection, based on the configuration of the first-phase Square Kilometre Array (SKA) low-frequency array. The resulted images are also compared with the result of the two-dimensional Fourier transform method. The results show that the image quality and correctness derived from both faceting and $w$-projection are better than the two-dimensional Fourier transform method in wide-field imaging. The effects of the number of facets and the $w$-direction step length on the image quality and running time are evaluated. The results indicate that the number of facets and the $w$-direction step length must be reasonable. Finally, we analyze the effect of data size on the running times of the faceting and $w$-projection algorithms. The results show that the faceting and $w$-projection algorithms need to be optimized before the huge amount of data processing. The research of the present paper initiates the analysis of wide-field imaging techniques and their applications in the existing and future low-frequency arrays, and will foster their applications in even broader fields.     

Imaging With The Mauritius Radio Telescope Challenges and Results

The Mauritius Radio Telescope (MRT) has been built with the main objective of surveying the southern sky at meter wavelengths. MRT is a Fourier synthesis, T-shaped non-coplanar array. It consists of a2048 m long East-West arm with 1024 fixed helices and a 880 m long South arm with 15 trolleys. Each trolley has four helices. A 512 channel, 2-bit 3-level complex correlation receiver is used to process the data from the EW and S group outputs. At least 60 days of observing are required for obtaining the Fourier components of the brightness distribution of the sky required to complete the survey. The MRT survey will be one of the most extensive survey at low frequencies providing a moderately deep radio catalog reaching a source density of about 2 × 10⁴ sr^-1 over most of the sky south of δ=-10^°with an angular resolution of 4' × 4.6' sec (δ) and a limiting flux density of 200 mJy (3 σ level) at 151 MHz. This paper will describe the telescope, the observations carried out so far, challenges of imaging with the data acquired over a period exceeding four years with a non-coplanar array, and summarises the results obtained so far. This revised version was published online in July 2006 with corrections to the Cover Date.     

基于OpenCL的射电干涉阵成像网格化算法实现

天文软件开发与应用中迫切需要在单机环境下进行高性能的科学数据处理,由于机器配置不同,采用传统的CUDA+GPU技术存在明显的局限,不利于天文软件的快速移植和无缝运行。针对明安图射电频谱日像仪数据处理中的网格化(Gridding)算法,采用并行计算OpenCL技术进行多线程编程实现。实验结果表明,基于OpenCL实现的网格化算法不仅能够在图形处理器上运行,而且能够在纯中央处理器上运行。当选择在图形处理器上执行时,算法的执行效率与基于CUDA实现的网格化算法执行效率大致相当,但算法不局限于NVIDIA GPU,解决了算法对CUDA+GPU的依赖;同时算法也能在纯中央处理器上较快速地执行,适用于单机模式下进行天文软件的开发和测试,也便于天文软件的应用与推广。     

Cost Effective Frequency Ranges For Multi-Beam Dishes, Cylinders, Aperture Arrays, and Hybrids

Five out of six Square Kilometre Array (SKA) science programs need extensive surveys at frequencies below 1.4 GHz and only four need high-frequency observations. The latter ones drive to expensive high surface accuracy collecting area, while the former ask for multi-beam receiver systems and extensive post correlation processing. In this paper, we analyze the system cost of a SKA when the field-of-view (Fov) is extended from 1 deg² at 1.4 GHz to 200 deg² at 0.7 GHz for three different antenna concepts. We start our analysis by discussing the fundamental limitations and cost issues of wide-band focal plane arrays (FPA) in dishes and cylinders and of wide-band receptors in aperture arrays. We will show that a hybrid SKA in three different antenna technologies will give the highest effective sensitivity for all six key science programs.     

Observing and Calibration Strategies for FIR Imaging with SIRTF

We present data from a 16×32 array of unstressed Ge:Gadetectors, showing the long time constant responses and hook effectcharacteristic of these devices. This array is half the size of the70 m imaging array planned for MIPS, the Multiband ImagingPhotometer for SIRTF. We present simulations of the FIR (farinfrared) sky as viewed by a MIPS array, including realistic noiseeffects and extraction of point sources with a standard photometrypackage, showing that the photometric goals of MIPS are attainablewith the current array performance and using standard analysistechniques. Previous calculations by Heim et al. (1998) haddetermined that MIPS would achieve 5- detections of 1 mJy at 70 m in 2000 s; the simulations presented hereobtain results consistent with those calculations.     

基于分布式执行框架的低频射电干涉阵列成像管线优化

平方公里阵列(Square Kilometre Array,SKA)项目是建设全球最大射电望远镜的国际合作项目,其灵敏度和测量速度将比当前所有的射电望远镜都要高出一个数量级.连续谱巡天是SKA的主要观测模式之一,基于连续谱成像建立巡天区域的标准星图,将能为后续天文科学研究奠定重要基础.银河系与河外星系全天默奇森宽场阵列拓展巡天(GaLactic and Extragalactic All-sky Murchison Widefield Array survey eXtended,GLEAM-X)是2018—2020年利用SKA先导望远镜默奇森宽场阵列(Murchison Wide-field Array,MWA)二期拓展阵列开展的新的射电连续谱巡天项目,观测期间积累了大量的低频巡天观测数据.海量观测数据的自动化、大批量处理是SKA望远镜项目所面临的的最大挑战和难题之一,基于分布式执行框架的成像管线优化经验将有助于解决海量数据处理问题.详细介绍了GLEAM-X成像管线并对其进行整合和改进,在中国SKA区域中心原型机(China SKA Regional Centre Prototype,...     

Doppler Imaging of Rapidly-Rotating M Stars

With the advent of 8–12m-class telescopes and powerful new spectrographs, we can now extend the Doppler-imaging technique to the cool (and faint) end of the main sequence. At a spectral type of approximately M2, stars are thought to become fully convective and cannot possess an overshoot layer between a radiative core and a convective envelope which, as in the case of the Sun and similar stars, likely harbors the dynamo. Therefore, one could expect a fundamentally different magnetic-field topology than on the Sun and thus a qualitatively different surface temperature distribution with new, hitherto unknown, magnetic activity phenomena. Unfortunately, most single M stars do not rotate sufficiently fast for Doppler imaging and one has to “use” binaries or pre-main-sequence stars in which M stars appear spun up or, in binaries, synchronized to the orbital motion.     

基于Dask并行加速的射电干涉成像网格化方法实现

快速傅里叶变换(Fast Fourier Transform,FFT)比傅里叶变换有更好的算法性能,是射电干涉成像的基础算法,但因为天线阵列的不规则采样,需使用网格化算法将可见度数据重采样到规则的网格上才能应用。基于卷积的网格化计算具有密集型和迭代型的特点,特别是处理海量可见度数据的情况下,高性能的网格化计算对整个成像过程加速尤为重要。为了缓解数据处理的压力,在现有处理整块数据和支持多核计算的算法基础上,拓展应用Dask并行计算框架,不仅将数据分块并分配到多线程上,提高数值计算效率,而且动态的分布式任务调度策略优化了网格化的实时处理。实验结果表明,多核中央处理器利用率显著提高,即使增加数据量,也能进一步提高网格化算法的性能。分布式任务调度能够将单(多)测量集的网格化弹性缩放到单(多)机系统,充分发挥集群的规模化优势。     

Imaging stars and their environments with the VLTI

The opportunity to deliver high-angular resolution model-independent images is one of the most attractive prospects for optical/infrared interferometry. In this paper I use simple imaging simulations to identify some of the practical difficulties that may arise in achieving this goal with the VLTI. For the types of studies investigated here – stellar surface imaging and the mapping of emission line disks – it is likely that the key challenges will be to measure the visibility function on a sufficiently large range of baselines, and to secure accurate and reliable phase information. This revised version was published online in July 2006 with corrections to the Cover Date.     

基于现场可编程门阵列的幸运成像算法的实现

幸运成像技术是用于消除天文图像中大气湍流影响的高分辨率图像重建技术,传统的基于中央处理器的幸运成像算法难于实时化。利用现场可编程门阵列并行处理的优势,设计了一种基于现场可编程门阵列的幸运成像算法并构建了一个实验系统。该系统用现场可编程门阵列完成了选图、配准、叠加的全部幸运成像算法流程,所得高分辨率图像与基于传统中央处理器算法处理的结果完全相同,但幸运成像算法的处理速度比传统中央处理器的处理速度快19倍。该算法在现场可编程门阵列上的实现,为幸运成像技术的实时或准实时化提供了一种可行的方案。     

Radio synthesis imaging of anisotropic angular broadening in the solar wind

We present Very Large Array observations at wavelengths of 2, 3.5, 6, and 20 cm, of angular broadening of radio sources due to the solar wind in the region 2–16 solar radii. Angular broadening is anisotropic with axial ratios in the range 2–16. Larger axial ratios are observed preferentially at smaller solar distances. Assuming that anisotropy is due to scattering blobs elongated along magnetic field lines, the distribution of position angles of the elliptically broadened images indicates that the field lines are non-radial even at the largest heliocentric distances observed here. At 5R⊙, the major axis scattering angle is ∼ 0.7" atλ= 6 cm and it varies with heliocentric distance asR ^-1.6. The level of turbulence, characterized by the wave structure function at a scale of 10 km along the major axis, normalized toλ = 20 cm, has a value 20 ± 7 at 5R⊙and varies with heliocentric distance asR ^-3. Comparison with earlier resu lts suggest that the level of turbulence is higher during solar maximum. Assuming a power-law spectrum of electron density fluctuations, the fitted spectral exponents have values in the range 2.8–3.4 for scale sizes between 2–35 km. The data suggests temporal fluctuations (of up to 10%) in the spectral exponent on a time scale of a few tens of minutes. The observed structure functions at different solar distances do not show any evidence for an inner scale; the upper limits are l k m at 2R⊙ and 4 km at 13R⊙. These upper limits are in conflict with earlier determinations and may suggest a reduced inner scale during solar maximum.     

Imaging diffuse emission with COMPTEL

The imaging telescope COMPTEL aboard the NASA Compton Observatory satellite has been demonstrated to be capable of imaging diffuse emission along the Galactic plane. Here we describe details of the imaging data spaces and methods to model the background. Different methods of imaging analysis are compared. Verification of consistency among these methods is a key factor in overall assessment of diffuse emission results. Source simulations and statistical analysis through the bootstrap method are applied to verify the significances of image structures.     

基于积分视场单元的三维天文成像光谱技术

近年来,三维成像光谱技术主要是发展和采用积分视场单元方法,即将枧场内的展源目标连续切割成若干单元,重新排列后进入光谱仪器,同时获得展源的高分辨率三维数据立方体(x,y和λ).相对于传统的技术方法,这种基于积分枧场单元的三维成像光谱技术进行一次观测可以同时获取天体各处的二维空间信息和光谱信息.采用积分视场单元虽然在光谱数据处理上会带来很多困难,但由于其在观测时间和效率上具有明显的优点,因此值得采用.该文简要介绍了三维成像光谱技术的原理;评述了目前实现三维成像光谱的三种不同类型的技术系统:小透镜阵系统、光纤加小透镜阵系统、像切分器系统,重点介绍了像切分器系统的有关情况;最后展望了三维成像光谱技术在天文学上的应用.