Improved Initialization of State-Space Artificial Neural Networks

被引:0
|
作者
Schoukens, Maarten [1 ]
机构
[1] Eindhoven Univ Technol, Dept Elect Engn, NL-5612 AZ Eindhoven, Netherlands
基金
欧盟地平线“2020”;
关键词
NONLINEAR-SYSTEM IDENTIFICATION; MODELS;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
The identification of black-box nonlinear state-space models requires a flexible representation of the state and output equation. Artificial neural networks have proven to provide such a representation. However, as in many identification problems, a nonlinear optimization problem needs to be solved to obtain the model parameters (layer weights and biases). A well-thought initialization of these model parameters can often avoid that the nonlinear optimization algorithm converges to a poorly performing local minimum of the considered cost function. This paper introduces an improved initialization approach for nonlinear state-space models represented as a recurrent artificial neural network and emphasizes the importance of including an explicit linear term in the model structure. Some of the neural network weights are initialized starting from a linear approximation of the nonlinear system, while others are initialized using random values or zeros. The effectiveness of the proposed initialization approach over previously proposed methods is illustrated on two benchmark examples.
引用
收藏
页码:1913 / 1918
页数:6
相关论文
共 50 条
  • [21] State-Space Analysis of Boolean Networks
    Cheng, Daizhan
    Qi, Hongsheng
    IEEE TRANSACTIONS ON NEURAL NETWORKS, 2010, 21 (04): : 584 - 594
  • [22] State-space control theory based analysis of feedforward neural networks.
    Craddock, RJ
    Warwick, K
    IEEE WORLD CONGRESS ON COMPUTATIONAL INTELLIGENCE, 1998, : 1383 - 1387
  • [23] Remodelling State-Space Prediction With Deep Neural Networks for Probabilistic Load Forecasting
    Arora, Parul
    Khosravi, Abbas
    Panigrahi, B. K.
    Suganthan, P. N.
    IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE, 2022, 6 (03): : 628 - 637
  • [24] State-space modeling for control based on physics-informed neural networks
    Arnold, Florian
    King, Rudibert
    ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE, 2021, 101
  • [25] Latent State-Space Models for Neural Decoding
    Aghagolzadeh, Mehdi
    Truccolo, Wilson
    2014 36TH ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY (EMBC), 2014, : 3033 - 3036
  • [26] Initialization Approach for Nonlinear State-Space Identification via the Subspace Encoder Approach
    Ramkannan, Rishi
    Beintema, Gerben I.
    Toth, Roland
    Schoukens, Maarten
    IFAC PAPERSONLINE, 2023, 56 (02): : 5146 - 5151
  • [27] Initialization of nonlinear state-space models applied to the Wiener-Hammerstein benchmark
    Marconato, Anna
    Sjoberg, Jonas
    Schoukens, Johan
    CONTROL ENGINEERING PRACTICE, 2012, 20 (11) : 1126 - 1132
  • [28] STATE-SPACE ANALYSIS OF ACTIVE RLC NETWORKS
    HUSEYIN, O
    ARCHIV FUR ELEKTRONIK UND UBERTRAGUNGSTECHNIK, 1973, 27 (02): : 91 - 96
  • [29] State-space modeling for degrading systems with stochastic neural networks and dynamic Bayesian layers
    Farhat, Md Tanzin
    Moghaddass, Ramin
    IISE TRANSACTIONS, 2024, 56 (05) : 497 - 514
  • [30] Predicting urban arterial travel time with state-space neural networks and Kalman filters
    Liu, Hao
    van Zuylen, Henk
    van Lint, Hans
    Salomons, Maria
    ARTIFICIAL INTELLIGENCE AND ADVANCED COMPUTING APPLICATIONS, 2006, (1968): : 99 - 108