Millisecond Exoplanet Imaging, II: Regression Equations and Technical Discussion

TL;DR

This paper discusses regression equations and technical details for simultaneously estimating non-common path aberrations and exoplanet images using millisecond telemetry, improving direct imaging of exoplanets amidst speckle noise.

Contribution

It provides detailed regression equations and analysis for three estimators, advancing the methodology for exoplanet imaging by accounting for wavefront sensor errors.

Findings

Regression equations for ideal, naive, and bias-corrected estimators.

Simulation results comparing estimator performances.

Method to incorporate wavefront sensor errors into exoplanet imaging.

Abstract

The leading difficulty in achieving the contrast necessary to directly image exoplanets and associated structures (eg. protoplanetary disks) at wavelengths ranging from the visible to the infrared are quasi-static speckles, and they are hard to distinguish from planets at the necessary level of precision. The source of the quasi-static speckles is hardware aberrations that are not compensated by the adaptive optics system. These aberrations are called non-common path aberrations (NCPA). In 2013, Frazin showed how, in principle, simultaneous millisecond (ms) telemetry from the wavefront sensor (WFS) and the science camera behind a stellar coronagraph can be used as input into a regression scheme that simultaneously and self-consistently estimates the NCPA and the sought-after image of the planetary system (the exoplanet image). The physical principle underlying the regression method is rather simple: the wavefronts, which are measured by the WFS, modulate the speckles caused by the NCPA and therefore can be used as probes of the optical system. The most important departure from realism in the author's 2013 article was the assumption that the WFS made error-free measurements. The simulations in Part I provide results on the joint regression on the NCPA and the exoplanet image from three different methods, called the ideal, the naive, and the bias-corrected estimators. The ideal estimator is not physically realizable but is a useful as a benchmark for simulation studies, but the other two are, at least in principle. This article provides the regression equations for all three of these estimators as well as a supporting technical discussion. Briefly, the naive estimator simply uses the noisy WFS measurements without any attempt to account for the errors, and the bias-corrected estimator uses statistical knowledge of the wavefronts to treat errors in the WFS measurements.