Application of BP Neural Network in PSD Nonlinear Correction

Article Preview

Abstract:

Because resistance of two-dimensional position sensitive detector's (PSD) photo surface is not absolute uniformity that its output is nonlinear. It is this feature enables the PSD difficult to measure small displacement. In order to solve this problem, BP neural network is proposed to solve the problem of PSD nonlinear correction after the study of traditional nonlinear correction method; BP neural network would have a strong ability of nonlinear mapping after training, and it can approach arbitrarily contact function by arbitrary precision, and MATLAB neural networking boxes can simulate BP neural network easily. Simulation and verification indicate that the method has a remarkable effect in solving nonlinear problems, and it can meet system requirements.

You might also be interested in these eBooks

Info:

Periodical:

Pages:

400-404

Citation:

Online since:

September 2014

Export:

Price:

* - Corresponding Author

[1] Yuan H X, Wang Z X, He A Z. Algorithm study of nonlinear correction of PSD[J]. Journal of Instrument Ap-pearance. 1999, (3)(in Chinese).

Google Scholar

[2] A. Makymen, Linear and sensitive CMOS position-sensitive photo detector, Electronics Letters, 1998, 34(12): 1255-1256.

Google Scholar

[3] LU Zhiyuan, ZHOU Yongjun, LI Wei jun. Study on method of nonlinear errors correction of sensor using BP neural network [J]. Journal of Transducer Technology,. Vol. 24 No. 2 2005 11-12.

Google Scholar

[4] LIU Jingang, HAN Xie, LI Hualing. Application of an Improved Arithmetic on BP Neural Networks [J]. Journal of North China Institute of Technology,. Vol. 23 No. 6 (2002).

Google Scholar

[5] GUI Xian- cai,Realization of BP Networks and Their Applications on MATLAB [J]. Journal of Zhanjiang Teachers College, Jun, 2004. Vol. 25 No. 3.

Google Scholar