Data streams from the low frequency instrument on-board the PLANCK satellite: Statistical analysis and compression efficiency

The expected data rate produced by the Low Frequency Instrument (LFI) planned to fly on the ESA PLANCK mission in 2007, is over a factor 8 larger than the bandwidth allowed by the spacecraft transmission system to download the LFI data. We discuss the application of lossless compression to PLANCK-LFI data streams in order to reduce the overall data flow. We perform both theoretical analysis and experimental tests using realistically simulated data streams in order to fix the statistical properties of the signal and the maximal compression rate allowed by several lossless compression algorithms. We studied the influence of signal composition and of acquisition parameters on the compression rate C-r and develop a semiempirical formalism to account for it. The best performing compressor tested up to now is the arithmetic compression of order 1, designed for optimizing the compression of white noise like signals, which allows an overall compression rate (C-r) over bar = 2.65 +/-0.02. We find that such result is not improved by other lossless compressors, being the signal almost white noise dominated. Lossless compression algorithms alone will not solve the bandwidth problem but needs to be combined with other techniques.