Efficient derivative-free Kalman filters for online learning.
作者： Eric A. Wan , Rudolph van der Merwe
DOI:
关键词:
摘要: The extended Kalman filter(EKF) is considered one of the most ef- fective methods for both nonlinear state estimation and parameter estimation(e.g., learning weights a neural network). Recently, number derivative free alternatives to EKF have been proposed. These include Unscented Filter(UKF) (1, 2), Central Difference Filter(CDF) (3) closely related Divided Filter(DDF) (4). filters consistently outperform estimation, at an equal computational complexity . Extension UKF was presented by Wan van der Merwe in (5, 6). In this paper, we further develop these techniques network training. extension CDF DDF their relation presented. Most significantly, paper introduces efficient square-root forms different filters. This enables implementation esti- mation (equivalent EKF), has added benefit improved numerical stability guaranteed positive semi-definiteness filter covariances.
参考文章(9)
E.A. Wan, R. Van Der Merwe, The unscented Kalman filter for nonlinear estimation Proceedings of the IEEE 2000 Adaptive Systems for Signal Processing, Communications, and Control Symposium (Cat. No.00EX373). pp. 0- 0 ,(2000) , 10.1109/ASSPCC.2000.882463
Simon J. Julier, Jeffrey K. Uhlmann, New extension of the Kalman filter to nonlinear systems Signal processing, sensor fusion, and target recognition. Conference. ,vol. 3068, pp. 182- 193 ,(1997) , 10.1117/12.280797
K. Ito, K. Xiong, Gaussian filters for nonlinear filtering problems IEEE Transactions on Automatic Control. ,vol. 45, pp. 910- 927 ,(2000) , 10.1109/9.855552
Lance Wu, Sharad Singhal, Training Multilayer Perceptrons with the Extended Kalman Algorithm neural information processing systems. ,vol. 1, pp. 133- 140 ,(1988)
Arnaud Doucet, Nando de Freitas, Eric A. Wan, Rudolph van der Merwe, The Unscented Particle Filter neural information processing systems. ,vol. 13, pp. 584- 590 ,(2000)
G.V. Puskorius, L.A. Feldkamp, Decoupled extended Kalman filter training of feedforward layered networks IJCNN-91-Seattle International Joint Conference on Neural Networks. ,vol. 1, pp. 771- 777 ,(1991) , 10.1109/IJCNN.1991.155276
S.J. Julier, J.K. Uhlmann, H.F. Durrant-Whyte, A new approach for filtering nonlinear systems advances in computing and communications. ,vol. 3, pp. 1628- 1632 ,(1995) , 10.1109/ACC.1995.529783
Alex T. Nelson, Eric A. Wan, Rudolph van der Merwe, Dual Estimation and the Unscented Transformation neural information processing systems. ,vol. 12, pp. 666- 672 ,(1999)
Niels Kjølstad Poulsen, Magnus Nørgaard, Ole Ravn, Advances in Derivative-Free State Estimation for Nonlinear Systems Informatics and Mathematical Modelling, Technical University of Denmark, DTU. ,(1998)