|
|
Adaptive algorithm for recursive identification of Hammerstein systems |
CHEN Kun1, LIU Yi1,2, WANG Hai-qing1, SONG Zhi-huan1, LI Ping1 |
(1. Institute of Industrial Process Control, State Key Laboratory of Industrial Control Technology, Zhejiang University, Hangzhou 310027, China;
2. Institute of Process Equipment and Control Engineering, Zhejiang University of Technology, Hangzhou 310032, China) |
|
|
Abstract A new on-line adaptive sparse, recursive identification algorithm of Hammerstein models based on output predictive error was proposed to solve the problems of the least squares support vector regression methods, such as lacking of sparsity and difficult to get a recursive form. The proposed method changes the recursive form, and adaptively chooses the strategy of sparseness/recursion/re-initialization according to the output predictive error. The error accumulation or even divergence problems are avoided and therefore the sparseness and accuracy are improved. The simulation illustrated that compared with the general recursive method, the proposed adaptive algorithm has sparse formulation and simplifies the model while keeping the identification accuracy. Also, the approach is robust and efficient, and it can meet the requirement of online identification.
|
Published: 26 February 2010
|
|
Hammerstein系统递推辨识的自适应算法
采用最小二乘支持向量回归对Hammerstein系统进行辨识缺乏稀疏性,且模型不易递推.提出一种基于输出预报误差的Hammerstein模型自适应稀疏递推辨识算法.根据分块矩阵对模型进行递推运算,基于系统输出预报误差的结果,自适应调整算法的辨识步骤,可以避免递推时可能出现的误差积累问题,有效提高算法的稀疏性和稳定性.仿真结果表明,与常规的递推算法相比,该自适应算法能够在保证辨识精度的情况下,有效稀疏和简化模型,提高算法的鲁棒性和辨识效率,更加符合系统在线辨识的需要.
|
|
[1] WESTWICK D T, KEARNEY R E. Identification of a Hammerstein model of the stretch reflex EMG using separable least squares [C]∥ Proceedings of the 22nd Annual International Engineering in Medicine and Biology Society. Chicago: IEEE, 2000, 3: 1901-1904.
[2] ESKINAT E, JOHNSON S H, LUYBEN W L. Use of Hammerstein models in identification of nonlinear systems [J]. AIChE Journal, 1991, 37(2): 255-268.
[3] PAWLAK M. On the series expansion approach to the identification [J]. IEEE Transactions on Automatic Control, 1991, 36(6): 736-767.
[4]DEMPSEY E J, WESTWICK D T. Identification of Hammerstein models with cubic spline nonlinearities [J]. IEEE Transactions on Biomedical Engineering, 2004, 51(2): 237-245.
[5] VAN PELT T H, BERNSTEIN D S. Nonlinear systems identification using Hammerstein and nonlinear feedback models with piecewise linear static maps [C]∥ Proceedings of American Control Conference. Chicago: IEEE, 2000, 1(6): 225-229.
[6] AL-DUWAISH H, KARIM M N. A new method for the identification of Hammerstein model [J]. Automatica, 1997, 33(10): 1871-1875.
[7] ESPINOZA M, SUYKENS J A K, DE MOOR B. Partially linear models and least squares support vector machines [C]∥ Proceedings of the 43rd IEEE Conference on Decision and Control. Atlantis, Bahamas: IEEE, 2004.
[8] GOETHALS I, PELCKMANS K, SUYKENS J A K, et al. Identification of MIMO Hammerstein models using least squares support vector machines [J]. Automatica, 2005, 41(7): 1263-1272.
[9] CHEN K, WANG H Q, SONG Z H. A recursive method of identification of Hammerstein model based on least squares support vector machines [C]∥ Proceedings of the 17th IFAC World Congress. Seoul: Elsevier, 2008: 4993-4998.
[10] LIU Y, YANG D C, WANG H Q, et al. Modeling of fermentation processes using online kernel learning algorithm [C]∥ Proceedings of the 17th IFAC World Congress. Seoul: Elsevier: 2008: 9679-9684.
[11] SUYKENS J A K, VAN GESTEL T, DE BRABANTER J, et al. Least squares support vector machines [M]. Singapore: World Scientific, 2002: 71-116.
[12] VAPNIK V. The nature of statistical learning theory [M]. New York: Springer, 1995: 181-218.
[13] WANG H Q, LI P, GAO F R, et al. Kernel classifier with adaptive structure and fixed memory for process diagnosis [J]. AIChE Journal, 2006, 52(10): 3515-3531. |
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
|
Shared |
|
|
|
|
|
Discussed |
|
|
|
|