Application of Improved Deep Neural Network in Complex Chemical Soft Measurement

Abstract

To improve the application effect of deep neural network in complex chemical soft measurement. After Kalman filtering equations were introduced, the algorithms of extended Kalman filter, volumetric Kalman filter and square root volumetric Kalman filter were optimized. After comparing the original algorithm with the optimization algorithm, we found that the design algorithm had better performance in all aspects. In this paper, the comparison between the original algorithm and the optimized algorithm is proved to be effective and the purpose of the algorithm optimization is achieved.