1: %URSI GA 2005  paper

2: %With Rose's comments

3: %With Prabu's comments

4: 

5: \documentclass[10pt]{article}

6: \usepackage{times}

7: \usepackage{graphicx}

8: 

9: \topmargin -0.5in

10: \oddsidemargin -0.12in

11: \evensidemargin -0.12in

12: \textheight 9.21 in

13: \textwidth 6.7in

14: \def\be{\begin{equation}}

15: \def\ee{\end{equation}}

16: \def\deg{$^{o}$}

17: \def\HI{HI}

18: \def\ie{i.e.}

19: \def\etal{et al}

20: 

21: 

22: \begin{document}

23: 

24: \begin{center}

25: {\Large\bf{

26: AN EXPERIMENTAL SETUP TO DEVELOP RFI MITIGATION TECHNIQUES

27: FOR RADIO ASTRONOMY}}

28: \vspace{3mm}

29: 

30: {\large\bf K. Jeeva Priya  and  D. Anish Roshi}

31: \vspace{3mm}

32: 

33: {\large\bf\itshape Raman Research Institute, Bangalore 560080 }

34: \vspace{3mm}

35: \end{center}

36: 

37: \noindent{\large\bf Abstract}

38: \vspace{3mm}

39: 

40: Increasing levels of Radio Frequency Interference (RFI) are a problem

41: for research in radio astronomy.  Various techniques to suppress

42: RFI and extract astronomical signals from data

43: affected by interference are being tried out. However, extracting

44: weak astronomical signals in the spectral region affected 

45: by RFI remains a technological

46: challenge. In this paper, we describe the construction

47: of an experimental setup at the Raman Research Institute (RRI),

48: Bangalore, India for research in RFI mitigation. We also present

49: some results of tests done on the data collected using this setup.

50: The experimental setup makes use of the 1.42 GHz receiver system

51: of the 10.4 m telescope at RRI. A new reference antenna, its receiver

52: system and a backend for recording digitized voltage together with

53: the 1.42 GHz receiver system form the experimental setup.

54: We present the results of the characterization of the experimental setup.

55: An off-line adaptive filter was successfully implemented and tested

56: using the data obtained with the experimental setup. \\

57: 

58: 

59: \noindent{\large\bf Introduction}

60: \vspace{3mm}

61: 

62: Radio frequency interference (RFI) is a growing problem for research

63: in radio astronomy. The ITU (International Telecommunication

64: Union) has defined and regulated the usage of the radio spectrum by

65: allocating various frequency bands for different

66: services including radio astronomy. Since radio

67: flux densities from cosmic sources are typically 40 to 100 dB below

68: those due to other services, often the out-of-band emission

69: from other services limits the sensitivity of astronomical

70: observations. Moreover, several radio spectral emissions

71: from cosmic sources are present in the frequency range

72: outside the allocated band for radio astronomy.

73: Thus mitigating interference and extracting

74: radio signals from cosmic sources is essential for advancement

75: of research in radio astronomy.

76: 

77: Several RFI mitigation techniques have been

78: developed and applied for radio astronomy application 

79: [1], [see also 2], [3 and references therein].

80: For example, a blanking algorithm to excise short (a few $\mu$sec wide) pulses

81: due to distance measuring equipment on astronomical data at frequencies

82: around 1 GHz was developed and applied to data taken using the Green Bank

83: telescope [4],[5]. A real-time adaptive cancellation technique to suppress

84: interference in the voltage domain was developed by [6], while [7]

85: developed a post-correlation technique, which essentially works on

86: the power domain (i.e. after detection). The deviation of the

87: probability distribution of the output of a

88: telescope from normal distribution in the presence of RFI was also used to

89: excise interference [8]. These techniques

90: are successful to a large extent, however, RFI rejection achieved

91: is not often sufficient for sensitive radio astronomy observations.

92: Thus research into developing new mitigation techniques is necessary. 

93: In this paper, we describe the construction of an experimental

94: setup at the Raman Research Institute (RRI) for developing

95: new mitigation techniques for radio astronomy application and present

96: some results of the test done on the data obtained with the setup. \\

97: 

98: \noindent{\large\bf The Experimental Setup}

99: \vspace{3mm}

100: 

101: 

102: %\begin{center}

103: \begin{figure}

104: \includegraphics[width = 6in,height = 3.5in,angle =0]{expsetup.eps}

105: %\includegraphics[width = 6in,height = 3.5in,angle =0]{expsetup.tif}

106: \caption{A block diagram of the experimental setup.}

107: \label{fig1}

108: \end{figure}

109: %\end{center}

110: 

111: A block diagram of the experimental setup is shown in Fig.~\ref{fig1}.

112: The 10.4 m telescope at RRI was originally built

113: for operation in the mm-wave band. A few years ago, a 1.4 GHz receiver

114: system was added to the telescope primarily for educational purposes.

115: The 1.4 GHz frequency band of this telescope is contaminated

116: with RFI and hence is an ideal system to develop and test interference

117: mitigation algorithms. We used one of the polarizations

118: of this receiver system as the `primary channel' to

119: receive the RFI contaminated astronomical signal.

120: We constructed a `reference channel' and used

121: the second polarization of the 1.4 GHz receiver system to

122: connect its output to the backend. The `reference channel' consists

123: of a pyramidal horn antenna, coupled to a low noise

124: amplifier (LNA) using a single linearized probe.

125: The horn antenna and the LNA operate near 1420.5 MHz.

126: The signal from the horn antenna is band limited to 25 MHz using a

127: cavity filter.

128: Both primary and reference channel outputs are further amplified

129: and down converted to 50 MHz in the interface unit. At this stage the

130: bandwidth is limited to 10 MHz. We modified an existing Portable

131: Pulsar receiver (PPR; [9]) to record the digitized voltage

132: from the two channels to a Personal Computer (PC) hard disk. The PPR

133: further limits

134: the bandwidth of the signals to 1.3 MHz. The

135: signals are bandpass sampled and digitized to 2 bits

136: before recording.

137: 

138: %\begin{center}

139: \begin{figure}

140: \includegraphics[width = 6in,height=3.5in,angle=0]{result1.eps}

141: %\includegraphics[width = 6in,height=3.5in,angle=0]{result1.tif}

142: \caption{Characteristics of the components of the experimental setup.

143: (a) E (black) and H (red) plane radiation patterns of the pyramidal horn

144: antenna. (b) Gain (top solid line marked 1) and output power (bottom solid

145: line marked 3) against input power to the LNA. The marker `c' indicates

146: 29 dB gain, which corresponds to the 1 dB compression point.

147: The 1 dB compression point of the LNA is $-$3.5 dBm (marker `c' in the bottom

148: curve). (c) Result of the two-tone, third order intercept point measurement

149: of the LNA. The output IP3 of the LNA is +12.5 dBm. (d) Result of the

150: measurement of the 1 dB compression point of the analog part of the

151: reference receiver system. The measured 1 dB compression point is

152: +17 dBm. }

153: \label{fig2}

154: \end{figure}

155: %\end{center}

156: 

157: The characteristics of the components of the experimental setup

158: were measured. The measured E and H plane radiation patterns of the pyramidal

159: horn antenna are shown in Fig.~\ref{fig2}. The half power beam

160: widths of the antenna in the E and H planes are 60$^o$ and 50$^o$

161: respectively. The LNA used in the `reference channel' is a

162: three stage HEMT amplifier [10], which has a measured gain of 30 dB,

163: noise temperature of 30 K and bandwidth of 500 MHz centered at 1.42 GHz.

164: The 1 dB compression point and the output IP3 (two-tone, third order

165: intercept point) of the LNA measured at 1.42 GHz are -3.5 dBm and

166: +12.5 dBm. We also measured the 1 dB compression point of the analog part

167: of the reference receiver system, which is +17 dBm.  \\

168: 

169: \noindent{\large\bf Preliminary Results}

170: \vspace{3mm}

171: 

172: %\begin{center}

173: \begin{figure}

174: \includegraphics[width = 6in,height=2.5in,angle=0]{result2.eps}

175: %\includegraphics[width = 6in,height=2.5in,angle=0]{result2.tif}

176: \caption{Spectra obtained from the data recorded using the

177: experimental setup. Spectra obtained from the reference and telescope

178: outputs are shown in (a) and (b) respectively. Two narrow band

179: interference can be seen in the plots. The Fourier transform

180: of the cross correlation of the two signals is shown in (c).

181: All spectra are integrated for about 60 msec.}

182: \label{fig3}

183: \end{figure}

184: %\end{center}

185: %\begin{center}

186: \begin{figure}

187: \includegraphics[width = 4in,height=1.5in,angle=0]{adfil.eps}

188: %\includegraphics[width = 4in,height=1.5in,angle=0]{adfil.tif}

189: \caption{A block diagram of the Matlab implementation of

190: the off-line adaptive filter. The 2 bit digitized data from

191: the experimental setup forms the inputs to the adaptive filter.

192: Fig.~\ref{fig5} shows the spectrum of the output of this filter.}

193: \label{fig4}

194: \end{figure}

195: %\end{center}

196: %\begin{center}

197: \begin{figure}

198: \includegraphics[width = 6in,height=2.5in,angle=0]{result3.eps}

199: %\includegraphics[width = 6in,height=2.5in,angle=0]{result3.tif}

200: \caption{Spectra of the telescope and reference antenna

201: outputs obtained with the radiated tone are shown in (a)

202: and (b) respectively. The spectrum of the output of the

203: adaptive filter is shown in (c). For comparison

204: the spectrum of the telescope output

205: is shown in `dash-dot' line in (c). As seen in the

206: figure the interference is attenuated after passing

207: through the adaptive filter. The radiated signal

208: (near normalized frequency of 0.5) is attenuated by

209: about 45 dB. }

210: \label{fig5}

211: \end{figure}

212: %\end{center}

213: 

214: We  recorded the data using the experimental system by pointing the

215: telescope at different directions. Sample spectra obtained from

216: the recorded data are shown in Fig.~\ref{fig3}. The signals are sampled

217: at a frequency slightly larger than the Nyquist rate (ie 1.3 $\times$ 2 MHz) for

218: the purpose of producing Fig.~\ref{fig3}. Two strong narrow band interferences

219: near 0.5 and 0.6 normalized frequencies are present at the

220: telescope and reference antenna outputs. The cross correlation of

221: the two outputs confirms the common origin of both these interferences.

222: 

223: For testing the functioning of the experimental setup, we radiated

224: a single tone and recorded the data from both the reference and

225: telescope output. The recorded signals were passed through

226: an adaptive filter configuration similar to that developed

227: by [6]. The block diagram

228: of the adaptive filter, which was implemented using Matlab,

229: is shown in Fig.~\ref{fig4}. A 60 tap FIR

230: filter along with least mean square (LMS) algorithm was

231: used for the implementation. The spectra of the input signals  

232: to the adaptive filter and the spectrum of its output are

233: shown in Fig.~\ref{fig5}. The spectra were integrated for

234: about 100 msec. The narrow band signal near the normalized frequency

235: 0.5 is the radiated tone. The narrow band signal near normalized

236: frequency 0.45 is an RFI from unknown source. As seen in Fig.~\ref{fig5},

237: both the radiated tone and the RFI are attenuated by the adaptive

238: filter. The RFI rejection obtained for the radiated tone is about 45 dB.

239: Note also that the spectral shape away from the narrow band

240: interferences is not

241: affected by the filtering process. \\

242: 

243: \noindent{\large\bf Acknowledgment}

244: \vspace{3mm}

245: 

246: We thank the staff of the radio astronomy laboratory and

247: mechanical workshop at RRI for their generous help during

248: the construction of the experimental setup. We

249: thank Prof. N. Udaya Shankar and Prof. C. R. Subrahmanya for

250: their support. The project described in this paper

251: is in partial fulfillment of KJP's M.E (Master of Engineering)

252: degree requirement. KJP thanks RRI for providing a project

253: assistantship. \\

254: 

255: \noindent{\large\bf Reference}

256: \vspace{3mm}

257: 

258: \noindent

259: [1]W. A. Baan, P. A. Fridman, R. P. Millenaar, ``Radio Frequency

260: Interference Mitigation at the Westerbork Synthesis Radio Telescope:

261: Algorithms, Test Observations, and System Implementation'', The

262: Astronomical Journal, Volume 128, pp. 933-949, 2004.

263: 

264: \noindent

265: [2]R. J. Fisher, ``RFI and How to Deal with It'', NAIC-NRAO School on

266: Single-Dish Radio Astronomy: Techniques and Applications, ASP conf.

267: series, eds. S. Stanimirovic, D. R. Altschuler, P. F. Goldsmith,

268: C. Salter, Vol. 278, p. 433-445, 2002.

269: 

270: \noindent

271: [3]P. A. Fridman, W. A. Baan, ``RFI mitigation methods in radio astronomy'',

272: Astronomy and Astrophysics, v.378, p.327-344, 2001.

273: 

274: \noindent

275: [4]J. R. Fisher, Q. Zhang, Y. Zheng, S. G. Wilson, R. F. Bradley,

276: ``Mitigation of Pulsed Interference to Redshifted H I and OH

277: Observations between 960 and 1215 Megahertz'',

278: The Astronomical Journal, Vol. 129, pp. 2940-2949, 2005.

279: 

280: \noindent

281: [5]Q. Zhang, Y. Zheng, S. G. Wilson, J. R. Fisher, R. F. Bradley,

282: ``Excision of Distance Measuring Equipment Interference from Radio

283: Astronomy Signals'',

284: The Astronomical Journal, Vol. 129, pp. 2933-2939, 2005.

285: 

286: \noindent

287: [6]C. Barnbaum, F. R. Bradley, ``A New Approach to Interference

288: Excision in Radio Astronomy: Real-Time Adaptive Cancellation'', The

289: Astronomical Journal, Volume 116, pp. 2598-2614, 1998.

290: 

291: \noindent

292: [7]F. H. Briggs, J. F. Bell, M. J. Kesteven,i ``Removing Radio

293: Interference from Contaminated Astronomical Spectra Using an

294: Independent Reference Signal and Closure Relations''

295: The Astronomical Journal, Volume 120, pp. 3351-3361, 2000.

296: 

297: \noindent

298: [8]P. A. Fridman, ``RFI excision using a higher order statistics

299: analysis of the power spectrum'', Astronomy and Astrophysics, vol.368,

300: pp. 369-376, 2001.

301: 

302: \noindent

303: [9]A. A. Deshpande et al., ``Portable Pulsar Receiver'', unpublished

304: 

305: \noindent

306: [10]A. Raghunathan, ``Building of the 21cm front end receiver for the

307: giant metrewave radio telescope'', M.Sc. Engg. Thesis, Bangalore

308: University, India, 2000

309: 

310: \end{document}

311: 

312: