Multivariable Optimization Method for Improving the Self-Calibration and Accuracy of a Binary Polarization State Analyzer

The binary polarization state analyzer (PSA) made of magnetooptic crystal (MO) rotators is a promising technology for achieving low-cost and high-efficiency state of polarization (SOP) analysis. To eliminate the influence of wavelength, temperature, and optical power noise on the internal optical components and data analysis of the PSA, a multivariable optimization method is proposed to solve the self-calibration problem and improve the SOP analysis accuracy. An overdetermined equation is constructed to describe the self-calibration problem of a 4-bit PSA, which consists of the rotation angle of the MO rotator, the retardation angle of the quarter-wave plate (QWP), and the relative orientation angle between the QWP and the polarizer (POL). Then, an optimization model is built using the least squares method in matrix form, and the optimization objective function is reasonably constructed by considering the optical power residual error and the degree of polarization (DOP). The trust region reflective algorithm is utilized to address the above model. Furthermore, the sensitivity of the proposed model is analyzed to choose the appropriate SOP for improving the self-calibration accuracy, and Monte Carlo analysis is executed to evaluate the accuracy of self-calibration and SOP analysis. Experimental results are consistent with simulation results. Compared with the method directly solving the linear equations of the 4-bit PSA, the presented method could achieve more accurate and self-calibrated measurements of the SOP.