K.J. Hunt, D. Sbarbaro, Neural Networks for Nonlinear Internal Model Control. IEE Proc.-D 138 (1991) 431-438.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... Furthermore, in most practical cases analytical models are not available, and therefore neural networks can be used to learn inverse dynamics, taking as input signal the outputs of the plant, and as supervision signal the inputs to the plant, either in the expected operational range of the plant [17], or in the whole operating space. FasArt neuro fuzzy system offers these two possibilities to obtain the control module: by inverting direct dynamics fuzzy rules. In this case, fuzzy rules inversion can be applied to rules extracted from FasArt plant model weights, as shown in [14] To build ....

K.J. Hunt and D. Sbarbaro, "Neural networks for nonlinear internal model control", IEE proceedings, vol. 138, no. 5, pp. 431-438, 1991.

....propagated backwards through the plant. In order to do this, the knowledge of the plant Jacobian, J, is required. The plant Jacobian is usually not easy to find. For minimum phase systems, Saerens and Soquet [20] suggested using only the sign of the plant Jacobian. In addition, Hunt and Sbarbaro [10] have shown that the sign of the plant Jacobian must be uniform in the entire operational region for an inverse plant model to exist at all. As in our case the plant Jacobian has the same sign as the gain of the P controller, the P controller control effort ) k e K y is directly proportional to ....

K.J. Hunt, D. Sbarbaro, Neural Networks for Nonlinear Internal Model Control. IEE Proc.-D 138 (1991) 431-438.

....neural based control strategies [11]#12, 12]#13, 13]#14. Examples are the direct inverse neuro control scheme [14]#15, the specialised inverse neuro control strategy [15]#16, the adaptive neuro control with backpropagation through time algorithm [16]#17, 17]#18, the internal model control [18]#19 and the optimal neuro control [19]#20. #12 #13 #14 #16 #17 #18 #19 #20 Another viable control strategy involves NN within a predictive control framework. According to this approach several works have been reported referring feedforward neural networks (FNN) with external recurrence and ....

Hunt, K. and Sbarbaro, D. - Neural networks for non-linear internal model control - IEE Proc D, 138, 431438,

....ffl The second problem concerns causality of the controller, which can be accomplished, when the reference signal is known temporal in advance. ffl The third problem is to avoid excessive control action, which is accomplished by a suitable low pass filtering of the reference signal. In [1] and [2] different control concepts based on a trained neural network, representing the inverse process model, are mentioned, e.g. Direct Inverse Control and Internal Model Control. In this paper, however, a trained neural network, representing the inverse process model, is applied in a control structure ....

....is described by a forward second order NARX model. Y (k 1) F (Y (k) Y (k Gamma 1) U(k) U(k Gamma 1) D(k) Y (k 1) Y (k 1) E(k 1) 4) where F is a non linear vector function, D(k) is the disturbance and E(k) is the output prediction error. In K.J. Hunt and D. Sbarbaro [2] it is proofed that, if F is monotonic with respect to U(k) then (4) is invertible, and the desired inverse process model is U(k) G (Y (k 1) Y (k) Y (k Gamma 1) U(k Gamma 1) D(k) 20 10 0 10 20 1 0.5 0 0.5 1 DISPLACEMENT corr(E1m,E1m) std.dev. 0.08034 [cm] 20 10 0 10 20 ....

K.J. Hunt and D. Sbarbaro. Neural networks for nonlinear internal model control. IEE ProceedingsD, Control theory and applications, 138(5):431--438, 1991.

.... model mismatch in the case of a linear model of the process [3] Developments of IM control in the case of nonlinear models of the process have been proposed, mainly for continous time models [4] but also for discrete time models [5] neural (discrete time) IM control systems are discussed in [6] [7]. Discrete time IM control characteristics are the consequence of the following properties: a) If the process and the controller are (input output) stable, and if the IM is perfect, then the control system is stable. b) If the process and the controller are stable, if the IM is perfect, if the ....

....than the process, due to a model mismatch and to disturbances. Remarks Neural IM control system are often simplified in the two following ways: Instead of the feedforward rallying model, the feedback reference model with input r and output z r is used to drive the inverse model, as in [7] for example. This assumes that z(k) z ral (k) k holds, i.e. that the inverse model is exact. But, dealing with general nonlinear neural models, this is usually not true. As a matter of fact, we showed in [13] that an IM control system with a rallying model is more robust towards a ....

Hunt K., Sbarbaro D. "Neural networks for nonlinear internal model control", IEE Proc.-D Vol. 138 N°5, 1991, pp.431-438.

....1.05 epoch mean squared error Simulation examples 63 6. 2 Model based neuro fuzzy controller During past several years, fuzzy control have emerged as one of the most active and fruitful areas for research in the applications of fuzzy logic [34] Neural networks have also been used as controllers [19]. The neuro fuzzy systems have been introduced to collect the strengths of neural networks and fuzzy logic. The neuro fuzzy systems are to capable of learning nonlinearities and the knowledge of the system is in a human understandable form as linguistic fuzzy rules. So far, the main emphasis in ....

K. J. Hunt and D. Sbarbaro, "Neural networks for nonlinear internal model control," IEE Proc.-D, vol. 138, no. 5, pp. 431--438, 1991.

No context found.

K.J. Hunt, D. Sbarbaro, Neural Networks for Nonlinear Internal Model Control. IEE Proc.-D 138 (1991) 431-438.

No context found.

Hunt K.J., Sbarbaro D. (1991). Neural networks for nonlinear internal model control, IEE Proceedings-D, Vol.138, No.5, pp.431-438.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC