Extended Kalman Filter Based on Generalized Regression Neural Network Simultaneous Determination of 2-Chlorophenol and 4-Chlorophenol

摘要

A spectrophotometry analysis relationship of nonlinear absorbance for binary mixture has been proposed. The nonlinear absorbance contained the correction factor that compensated the distortion by interactive effect of mixture. Using nonlinear Kalman filter UV spectrophotometric simultaneous determinate mixture of 2-chlorophenol and 4-chlorophenol firstly. Calibration set with 30 standard solutions (range 1-14 mg/L) and 61 wavelengths (260-290m, 0.5nm slit width). Standard coefficient matrix of extended Kalman filter was performed from generalized regression neural network. The vectors function Jacobian matrix were obtained by linearized from Taylor series for nonlinear absorbance formula. The recovery experiments showed the extended Kalman filter simultaneous determination of mixture for 2-chlorophenol and 4-chlorophenol is more accurate than CLS, PLS and linear Kalman filter, at same time, the EKF is a stable recursion method.