Derivatives of optical path length: from mathematical formulation to applications

Abstract

The optical path length (OPL) of an optical system is a highly important parameter since it determines the phase of the light passing through the system and governs the interference and diffraction of the rays as they propagate. The Jacobian and Hessian matrices of the OPL are of fundamental importance in tuning the performance of a system. However, the OPL varies as a recursive function of the incoming ray and the boundary variable vector, and hence computing the Jacobian and Hessian matrices is extremely challenging. In an earlier study by the present group, this problem was addressed by deriving the Jacobian matrix of the OPL with respect to all of the independent system variables of a nonaxially symmetric system. In the present study, the proposed method is extended to the Hessian matrix of a nonaxially symmetric optical system. The proposed method facilitates the cross-sensitivity analysis of the OPL with respect to arbitrary system variables and provides an ideal basis for automatic optical system design applications in which the merit function is defined in terms of wavefront aberrations. An illustrative example is given. It is shown that the proposed method requires fewer iterations than that based on the Jacobian matrix and yields a more reliable and precise optimization performance.

Design and Characterisation of a Fast Steering Mirror Compensation System Based on Double Porro Prisms by a Screw-Ray Tracing Method

“…The flat-boundary skew-ray tracing method, proposed by Lin [ 18 , 22 ], is used in this study to build the mathematical algorithm of the proposed FSM compensation system. The transformation matrix is relative to the universal coordinate system ( R ) which is built on the boundary of each optical element, as shown below: …”

Sequential differential ray tracing formulation for homogeneous and gradient-index lenses using automatic differentiation

“…13 DRT has been an active area of research in recent years; early works on DRT were limited to simple lens systems where constraints, such as symmetry, surface type, etc., were enforced to derive closed form analytical expressions. 14,15 The method was later extended 16,17 for more general and complex systems with no assumptions on lens geometry. Unfortunately, in all previously proposed DRT methods, the analytical expressions needed for evaluating the gradient of the merit function are derived by hand, leading to inconvenient and complicated equations not suitable for use in an optical ray tracer, which also limits the technique's generality.…”