Performances of CAA Algorithms for Alignment State Estimation Simulations for Three-mirror Anastigmat Earth Observation Optical System

Abstract

We report the performance study for several computer aided alignment algorithms for precision alignment state estimation of Three-Mirror Anastigmat(TMA) optical system that has been a popular choice for space optical cameras. They include sensitivity table method, merit function regression, differential wavefront sampling, and multiple design configuration optimization. First, we performed the sensitivity analysis and identified the tertiary mirror(M3) as the compensator. Its motion in two degrees of freedom(tilt X, tilt Y) are used to compensate the wavefront error because of the spherical mirror characteristic of M3. We simulated the alignment state estimation for the random misalignment to primary and secondary mirrors within the range of assembly tolerance(decenter X, Y: ±0.1mm, tilt X, Y: ±0.05 degree) while including hexapod alignment system error. The trial simulation result using merit function regression showed that we can meet the alignment requirement (1/15λ RMS wavefront error at 633nm) without iterations when random misalignments are not imposed to primary and secondary mirrors. The details of the computation, the results and implications are discussed.ensitivity table method, merit function regression, differential wavefront sampling, and multiple design configuration optimization. First, we performed the sensitivity analysis and identified the tertiary mirror(M3) as the compensator. Its motion in two degrees of freedom(tilt X, tilt Y) are used to compensate the wavefront error because of the spherical mirror characteristic of M3. We simulated the alignment state estimation for the random misalignment to primary and secondary mirrors within the range of assembly tolerance(decenter X, Y: ±0.1mm, tilt X, Y: ±0.05 degree) while including hexapod alignment system error. The trial simulation result using merit function regression showed that we can meet the alignment requirement (1/15λ RMS wavefront error at 633nm) without iterations when random misalignments are not imposed to primary and secondary mirrors. The details of the computation, the results and implications are discussed.