The most important aspect of differential neural networks dynamics is related to their weights properties. This is a consequence of the complex non-linear structure describing the learning matrix differential equations, which are associated with adaptive capability of this kind of neural network. So far, there is no analytical demonstration about the weights stability. In fact, this is the main inconvenience in designing real applications for differential neural network observers. This study deals with the stability proof for the weights dynamics using an adaptive procedure to adjust the weights ordinary differential equations. Three different examples (two of them were realised by numerical simulations and the last one was carried out using real biofiltering process data) demonstrated the good performance of the suggested approach.
Inspec keywords: observers; neural nets; matrix algebra; stability; differential equations
Other keywords:
Subjects: Simulation, modelling and identification; Linear algebra (numerical analysis); Differential equations (numerical analysis); Neural nets (theory)
References
-
-
1)
- W. Yu , M. Moreno , X. Li . Observer-based neuro identifier. IEE Proc. Control Theory Appl. , 145 - 152
-
2)
- V. Vapnik . (1995) The nature of statistical learning theory.
-
3)
- Derdiyok, A., Yan, Z., Guven, M., Utkin, V.: `A sliding mode speed and rotor time constant observer for induction machines', IECON'01: The 27th Annual Conf. of the IEEE Industrial Electronics Society, 2001.
-
4)
- S. Haykin . (1999) Neural networks. A comprehensive foundation.
-
5)
- D. Luenberger . Obseving the state of linear sistems. IEEE Trans. Military Electron. , 74 - 90
-
6)
- R.P.N. Rao . Bayesian computation in recurrent neural circuits. Neural Comput. , 1 , 1 - 38
-
7)
- J. Raol . Parameter estimation of state space models by recurrent neural networks. IEE Proc. Control Theory Appl. , 2 , 114 - 118
-
8)
- M. Vidyasagar . Randomized algorithms for robust controller synthesis using statistical learning theory. Automatica , 10 , 1515 - 1528
-
9)
- A. Meghdari , D. Naderi , M.R. Alam . Neural network based observer for real time tipover estimation. Mechatronics , 989 - 1004
-
10)
- A. Levant . Sliding order and sliding accuracy in sliding mode control. Int. J. Control , 6 , 1247 - 1263
-
11)
- J.P. Gauthier , G. Bornard . Observability for any u(t) of a class of nonlinear systems. IEEE Trans. Aut. Control , 922 - 926
-
12)
- A.S. Poznyak , J. Medel . Matrix forgetting with adaptation. Int. Journal of Systems Science , 8 , 865 - 878
-
13)
- A.S. Poznyak , J. Medel . Matrix forgetting factor. Int. J. Systems Science , 2 , 165 - 174
-
14)
- Hovakimyan, N., Rysdyk, R., Calise, J.: `Dynamic neural networks for ooutput feedback control', Technical, 2004, School of Aerospace Engineering Georgia Institute of Technology.
-
15)
- A. Poznyak , E. Sanchez , W. Yu . (2001) Differential neural networks for robust nonlinear control (identification, state estimation an trajectory tracking).
-
16)
- K. Narendra , K. Parthasarathy . Identification and control of dynamical systems using neural networks. IEEE Trans. Neural Netw. , 4 - 27
-
17)
- G. Park , Z. Bien . Neural network based fuzzy observer with application to facial analysis. Pattern Recogn. Lett. , 93 - 105
-
18)
- A.F. Fillipov . (1998) Differential equations with disscontinuos right-hand side.
-
19)
- T. Poznyak , I. Chairez , A. Poznyak . Application of a neural observer to phenols ozonation in water: simulation and kinetic parameters identification. Water Res. , 2611 - 2620
-
20)
- A. Krener , A. Isidori . Linealization by output injection and nonlinear observers. Systems Control Lett. , 47 - 52
-
21)
- F. Abdollahi , H.A. Talebi , R.V. Patel . A stable neural network based observer with applications to flexible-joint manipulators. IEEE Trans. Neural Netw. , 118 - 129
-
22)
- J.J. Slotine , J.K. Hedrik , E.A. Misava . On sliding observers for nonlinear systems. ASME J. Dyn. Syst., Meas. Control , 245 - 252
-
23)
- Poznyak, A., Escobar, J., Shtessel, Y.: `Stochastic sliding modes identification', Proc. 2006 Workshop on Variable Structure Systems, June 2006, Alghero, Italy, p. 5–9.
-
24)
- Haiqing, W., Congwei, L., Fahai, L., Lixiang, W.: `A new flux observer of induction motor by hinf optimization', 36thIAS Annual Meeting. Conference Record on Industry Applications Conference, 2001.
-
25)
- J. Davila , L. Fridman , A. Levant . Second-order sliding-mode observer for mechanical systems. IEEE Trans. Autom. Control , 1785 - 1789
-
26)
- W.T. Miller , R.S. Sutton , P.J. Werbos . (1990) Neural networks for control.
-
27)
- U.T. Eden , L.M. Frank , V. Solo , E.N. Brown . Dynamic analysis of neural encoding by point process adaptive filtering. Neural Comput. , 971 - 998
-
28)
- J. Tani . Model-based learning for mobile robot navigation from the dynamical system perspective. IEEE Trans. Syst., Man Cybern. , 3 , 421 - 436
-
29)
- F.L. Lewis , A. Yesildirek , K. Liu . Multilayer neural-net robot controller with guaranteed tracking performance. IEEE Trans. Neural Netw. , 2 , 1 - 11
-
30)
- G. Rovithakis , M. Christodoulou . Adaptive control of unknown plants using dynamical neural networks. IEEE Trans. Syst., Man Cyben. , 400 - 412
-
31)
- Y.S. Boutalis , O. Kosmidou . A feedback linearization technique by using neural networks: application to bioprocess control. J. Comput. Methods Sci. Eng. , 269 - 282
-
32)
- B. Jayanand , V. Subrahmanyam . Sensorlees direct vector control of induction motor using a kalman filter trained neuro observer. Int. J. Eng. Intell. Syst. Electr. Eng. Commun. , 1 , 57 - 64
-
33)
- L. Jingchuan , X. Longya , Z. Zheng . An adaptive sliding-mode observer for induction motor sensorless speed control. IEEE Trans. Ind. Appl. , 1039 - 1046
-
34)
- P.J. Antsaklis . Guest editor's introduction. IEEE Control Syst. Mag. , 5 - 7
-
1)
Related content