Dynamic Measurement System Modeling with Machine Learning Algorithms

Authors: Changqiao Wu, Guoqing Ding, Xin Chen

Abstract:

In this paper, ways of modeling dynamic measurement systems are discussed. Specially, for linear system with single-input single-output, it could be modeled with shallow neural network. Then, gradient based optimization algorithms are used for searching the proper coefficients. Besides, method with normal equation and second order gradient descent are proposed to accelerate the modeling process, and ways of better gradient estimation are discussed. It shows that the mathematical essence of the learning objective is maximum likelihood with noises under Gaussian distribution. For conventional gradient descent, the mini-batch learning and gradient with momentum contribute to faster convergence and enhance model ability. Lastly, experimental results proved the effectiveness of second order gradient descent algorithm, and indicated that optimization with normal equation was the most suitable for linear dynamic models.

Keywords: Dynamic system modeling, neural network, normal equation, second order gradient descent.

Digital Object Identifier (DOI): doi.org/10.5281/zenodo.2022017

References:

[1] J. P. Xiang, “Dynamic properties of force transducer,” Process Automation instrumentation, no. 6, pp. 12–21+102, 1981.
[2] A. Simpkins, “System identification: Theory for the user, 2nd edition,” IEEE Robotics Automation Magazine, vol. 19, no. 2, pp. 95–96, June 2012.
[3] K. J. Xu and M. Yin, “A dynamic modeling method based on flann for wrist force sensor,” Chinese Journal of Scientific Instrument, vol. 21, no. 1, pp. 92–94, 2000.
[4] S. P. Tian, P. P. Jiang, and G. Z. Yan, “Application o-f recurrent neural network to dynamic modeling of sensors,” Chinese Journal of Scientific Instrument, vol. 25, no. 5, pp. 574–576, 2004.
[5] X. D. Wang, C. J. Zhang, and H. R. Zhang, “Sensor dynamic modeling using least square support vector machines,” Chinese Journal of Scientific Instrument, vol. 27, no. 7, pp. 730–733, 2006.
[6] D. H. Wu, W. Zhao, S. L. Huang, and S. K. He, “Research on improved flann for sensor dynamic modeling,” Chinese Journal of Scientific Instrument, no. 2, pp. 362–367, 2009.
[7] W. J. Yang, “Research on dynamic characteristics and compensation technology of pressure sensors,” Master’s thesis, North University of China, Shanxi, 2017.
[8] T. Dab´oczi, “Uncertainty of signal reconstruction in the case of jittery and noisy measurements,” IEEE Transactions on Instrumentation and Measurement, vol. 47, no. 5, pp. 1062–1066, 1998.
[9] J. Brignell, “Software techniques for sensor compensation,” Sensors and Actuators A: Physical, vol. 25, no. 1-3, pp. 29–35, 1990.
[10] M. A. Nielsen, Neural Networks and Deep Learning. Determination Press, 2015.
[11] S. P. Tian, Y. Zhao, W. H. Yu, and Z. W. Wang, “Nonlinear compensation of sensors based on bp neural network,” Journal of Test and Measurement Technology, vol. 21, no. 1, pp. 84–89, 2007.
[12] She Ping Tian, “Nonlinear dynamic compensation of sensors based on recurrent neural network model,” Journal of Shanghai Jiao Tong University, vol. 37, no. 1, pp. 13–16, 2003.
[13] H. M. Huang, “Dynamical compensation method for weighting sensor based on flann,” Transducer and Microsystem Technologies, vol. 25, no. 8, pp. 25–28, 2006.
[14] L. Q. Hou, W. G. Tong, and T. X. He, “Nonlinear errors correcting method of sensors based on rbf neural network,” Journal of Transducer Technology, vol. 17, no. 4, pp. 643–646, 2004.
[15] Y. LeCun, B. Boser, J. S. Denker, D. Henderson, R. E. Howard, W. Hubbard, and L. D. Jackel, “Backpropagation applied to handwritten zip code recognition,” Neural computation, vol. 1, no. 4, pp. 541–551, 1989.
[16] H. Li, Statistical learning method. Beijing: Tsing Hua University Press, 2012.
[17] D. H.Wu, “Dynamic compensating method for transducer based on flann inverse system constructed by ls-svm,” Journal of Data Acquisition and Processing, vol. 22, no. 3, pp. 378–383, 2007.
[18] I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. MIT Press, 2016.
[19] N. Qian, “On the momentum term in gradient descent learning algorithms,” Neural Networks, vol. 12, no. 1, pp. 145–151, 1 1999.