MAOS
Multithreaded Adaptive Optics Simulator
Loading...
Searching...
No Matches
Algorithms

This section describes some of the models and numerical algorithms used in MAOS.

DM Actuator Influence Function

Linear influence function

The linear influence function is defined as

\[h(x;x_{i};\delta)=\begin{cases} 1-|x-x_{i}|/\delta & \textrm{if }|x-x_{i}|\leq\delta\\ 0 & \textrm{else} \end{cases} \]

where \(\delta\) is usually equal to the grid spacing. A bilinear influence function is simply the product of two linear influence functions with variables \(x_{1}\) and \(x_{2}\).

Cubic influence function

A cubic influence function that can reproduce piston/tip/tilt was proposed by Ellerbroek to model the piezo stack DM actuator. The influence function has the form

\[h(x;x_{i};\delta)=h_{0}((x-x_{i})/\delta) \]

where \(\delta\) is the same as the grid spacing, and \(h_{0}\) is the influence function defined in the normalized coordinates

\[h_{0}(x)=\frac{1}{1+2c}\begin{cases} 1+(4c-\frac{5}{2})|x|^{2}+(\frac{3}{2}-3c)|x|^{3} & |x|\leq1\\ (2c-\frac{1}{2})(2-|x|)^{2}+(\frac{1}{2}-c)(2-|x|)^{3} & 1<|x|\leq2\\ 0 & |x|>2 \end{cases} \]

where \(c\) is the inter-actuator coupling coefficient: the ratio \(h_0(1)/h_0(0)\) of the response at the nearest neighbor actuator to that at the actuator itself. The leading coefficient normalizes the influence function so that the sum over the actuator grid points, \(h_0(0)+2h_0(1)\), equals 1. The 2-d influence function is the product of the influence functions along \(x\) and \(y\).

Implementation

The influence function is selected per grid by the iac member of loc_t and map_t: 0 uses the bilinear influence function, and a positive value uses the cubic influence function with \(c=\)iac.

  • Ray tracing from a grid or point set with iac set uses the cubic interpolation in lib/accphi.c (prop_grid_cubic(), prop_nongrid_cubic(), and related functions, called by prop()). Transpose ray tracing of a map_t is not available with the cubic influence function.
  • The sparse ray tracing operators built by mkh() and mkht() (lib/mkh.c) use mkht_cubic() automatically when the input loc_t has iac set.
  • The coefficients \(c_0\) to \(c_4\) of the two polynomial pieces are computed in the macro PREP_CUBIC_PARAM (lib/accphi.c) and in mkht_cubic().

The coupling is set by the following parameters:

  • dm.iac: the DM actuators (default 0.223 in sim.conf; 0 uses the bilinear influence function). It is assigned to recon->aloc.
  • moao.iac: the MOAO DM actuators.
  • tomo.iac: the tomography grid xloc (default 0, for testing).
  • dbg.wfs_iac: the grid used to fit turbulence onto the WFS subapertures (default 0). It is a real number like the others.