|
During days-long exposure of the Roman Space Telescope (RST) Coronographic Instrument (CGI), the starlight wavefront will drift due to thermal instabilities of the optical tube assembly (OTA) and deformable mirror (DM) actuators. While low-order spatial wavefront modes will be sensed upstream of the coronagraph, focal-plane wavefront sensing (FPWS) via DM probing will be used to correct high-order modes. In RST’s nominal operation mode, the high-contrast region of the image (dark hole) will be periodically retouched by pointing at a bright reference star, performing FPWS, correcting the wavefront, and pointing back at the target. We propose instead to continuously probe the DM during the observation, so that the wavefront can be estimated without slewing the telescope. This approach will reduce the risk of introducing OTA instabilities at the expense of slight worsening of the contrast due to the introduction of small DM probes. This work expands previous FPWS methods from single-pixel Extended Kalman Filters (EKF) to an EKF in terms of multiple DM actuators simultaneously. We then evaluate this approach on a numerical model of RST-CGI, based on Observing Scenario (OS) 9. Besides wavefront drift, we also simulate the effects of detector noise associated with photon-counting and line-of-sight jitter. Our results show that it is possible to correct the DM once every 24 hours, without slewing the telescope. This continuous FPWS is compared to periodic slewing based on the planet detection performance in post-processing. In particular, we report similar post-processing errors when doing Angular Differential Imaging after the nominal OS, and Electric Field Order Reduction after continuous dark hole maintenance.
|
