Neural network based surface shape modeling of stressed lap optical polishing

It is crucially important to establish an accurate model to represent the relationship between the actuator forces and the lap surface changes when polishing a large and highly aspheric optical surface. To facilitate a computer-controlled optical polishing process, a neural network based stressed lap surface shape model was developed. The developed model reflects the dynamic deformation of a stressed lap. The original data from the microdisplacement sensor matrix were used to train the neural network model. The experimental results show that the proposed model can represent the surface shape of the stressed lap accurately and provide an analytical model to be used to polish the stressed lap control system and the active support system for a large mirror. (C) 2010 Optical Society of America