Design of the on-board data compression for the bolometer data of LiteBIRD

Abstract: LiteBIRD is a space-borne experiment dedicated to detecting large-scale $B$-mode anisotropies in the linear polarization of the Cosmic Microwave Background (CMB) predicted by the theory of inflation. It is planned to be launched in the late 2020s to the second Lagrange point (L2) of the Sun-Earth system. LiteBIRD will map the sky in 15 frequency bands. In comparison to $\it{Planck}$ HFI, the previous low-temperature bolometer-based satellite for CMB observations, the number of detector has increased by two orders of magnitude, up to $\sim$5000 detectors in total. The data rate is 19 Hz from each detector. The bandpass to the ground is limited to 10 Mbps using the X-band for a few hours per day. These require the data to be compressed by more than 50 %. The exact value depends on how much information entropy is contained in the real data. We have thus evaluated the compression by simulating the time-ordered data of polarization sensitive bolometers. The foreground emission, detector noise, cosmic ray glitches, leakage from the CMB intensity to polarization, etc. are simulated. We investigated several algorithms and demonstrated that the required compression ratio can be achieved by some of them. We describe the details of this evaluation and propose algorithms that can be employed in the on-board digital electronics of LiteBIRD.