Simulated Annealing for Control of Adaptive Optics System

Many optical systems, such as imaging systems or laser communication systems, suffer performance degradation due to distortions in the optical wave-front. An optical wave propagates through an optically inhomogeneous medium such as the atmosphere, differences in the index of refraction along the propagation path cause variations in the speed of light propagation, which lead to phase distortions. Adaptive Optics (AO) techniques are often used to compensate these static or dynamic aberrations of a light beam after propagation through a distorting medium (Hardy, 1998). Although originally proposed for astronomical telescopes in 1953 (Babcock, 1953), adaptive optics did not become a reality until the 1970s, when it was developed for national defence applications, specifically laser beam compensation and satellite imaging. It consists of using an active optical element such as a deformable mirror to correct the instantaneous wavefront distortions. These are measured by a device called a wavefront sensor which delivers the signals necessary to drive the correcting element. The first adaptive optics system able to sharpen two-dimensional images was built at Itek by Hardy and his co-workers (Hardy et al, 1977). AO provides a means to perform real-time correction of aberrations imposed on light waves as they travel from the source to the imaging system. While AO has its roots the field of astronomy it is currently used in a wide variety of medical, military and industrial applications. The papers by Milonniand (Milonni, 1999) and Parenti (Parenti, 1992) provide an excellent introduction to the use of AO in Astronomy. A comprehensive review of the medical and industrial applications of AO can be found in the technology tracking report by Greenaway and Burnett (Greenaway, & Burnett, 2004). The most common conventional adaptive optics systems (Fig. 1) is implemented with a wave-front corrector to correct the distorted wave-front, a wave-front sensor to measure the aberrations present in the incoming beam, and a feedback control algorithm to link these two elements in real time. Although the technique based on rapid wave-front measurement has been found useful in astronomical applications, this approach to the control problem is much difficult to be used in situations where wave-front distortions can not be measured 15