Influence of regularization methods on the operation quality of neural state estimators of two-mass drive system

被引：0

作者：

Orlowska-Kowalska, Teresa ^{[1
]}

Kaminski, Marcin ^{[1
]}

机构：

[1] Wroclaw Univ Technol, Inst Maszyn Napedow & Pomiarow Elekt, PL-50372 Wroclaw, Poland

来源：

PRZEGLAD ELEKTROTECHNICZNY | 2010年 / 86卷 / 04期

关键词：

neural networks; training methods; regularization; two-mass system; state variable estimation; TIKHONOV REGULARIZATION; NETWORKS;

D O I：

暂无

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

In the paper the influence of training strategies of neural networks, using different regularization methods, on the state variable estimation of the drive system with elastic joint is demonstrated. The most frequently used regularization strategies are taking into account. The influence of specific parameters of these methods on the estimation quality of torsional torque and load side speed of the two-mass drive is tested. Moreover the robustness of the obtained neural estimators to the load inertia changes is tested. (Influence of regularization methods on the operation quality of neural state estimators of two-mass drive system).

引用

页码：174 / 179

页数：6

共 18 条

[1]

[Anonymous], 1977, NEW YORK

[2]

[Anonymous], 1963, Soviet Math

[3]

BELICZYNSKI B, 2000, PRACE NAUKOWE PW E, V112

[4] TRAINING WITH NOISE IS EQUIVALENT TO TIKHONOV REGULARIZATION [J].

BISHOP, CM .

NEURAL COMPUTATION, 1995, 7 (01) :108-116

[5]

BOS S, 1996, IEEE INT C NEURAL NE, V1, P241

[6] Analysis of Tikhonov regularization for function approximation by neural networks [J].

Burger, M ;

Neubauer, A .

NEURAL NETWORKS, 2003, 16 (01) :79-90

[7]

CHAN L, 1996, PROGR NEURAL INFORM, P139

[8]

Ganatsos T., 2007, PIERS ONLINE, V3, P1165

[9] REGULARIZATION THEORY AND NEURAL NETWORKS ARCHITECTURES [J].

GIROSI, F ;

JONES, M ;

POGGIO, T .

NEURAL COMPUTATION, 1995, 7 (02) :219-269

[10]

GOUTTE C, 1998, IEEE P INT C ACOUSTI, V2, P1201

← 1 2 →