The algorithm used to identify wind in Adaptive Optics telemetry data.
Adaptive Optics is a technique used in astronomical instrumentation to correct for turbulence in the earth’s atmosphere in real time. Atmospheric turbulence tends to blur out images taken at telescopes. To counteract this blurring, astronomers must either take very short images (which would limit our ability to detect very faint objects) or correct for this turbulence in real time.
Adaptive optics (AO) systems correct for this turbulence in real time. AO systems use a deformable mirror, connected to many actuators which control the shape of the front surface of the mirror thousands of times per second. Using a separate sensor, AO systems measure the shape of incoming atmospheric turbulence (called the wavefront).
A significant source of error and difficulty in AO systems is the computational delay between measuring the incoming wavefront and moving the deformable mirror to correct the turbulence. Using new techniques, developed with Lisa Poyneer, we can eliminate the effects of the delay. Our technique revolves around predicting what the atmospheric turbulence will be doing when we apply our correction.
To predict atmospheric turbulence, we search for wind velocities in the atmospheric column above the telescope. Using these wind velocities, we can translate the observed turbulence across the telescope aperture. We perform this translation using a Kalman Filter which allows us to translate and predict the effects of several wind layers simultaneously.