Neural network controller

Neural network control systems 10.3.1 Artificial neural networks... 

M. Karpenko, N. Sepehri, and D. Scuse. Diagnosis of process valve actuator faults using a multilayer neural network. Control Engineering Practice, 11 1289-1299, 2003. 

D.L. Yu, J.B. Gomm, and D. Williams. Sensor fault diagnosis in a chemical process via RBF neural networks. Control Engineering Practice, 7 49-55, 1999. 

Model-Free Adaptive and Artificial Neural Network Control... 

This chapter is intended to provide an introduction to neural network techniques and a guide to their application to biomedical control systems problems. The reader is referred to recently published textbooks and numerous journal articles for the specific information required to implement a given neural network control system. 

In designing a neural network control system, one must select the overall structure of the system and decide which components will utilize neural network algorithms. Several examples of control system structures are provided below, each of which utilizes one or more neural networks as described above. This section of the chapter provides a brief overview of some neural control systems that have potential for application in biomedical control systems. For excellent, thorough reviews of recent developments in neural network control systems, the reader is referred to Miller [1990b] and White and Sofge [ 1992]. 

Neural network approaches have been used as an alternative to other nonhnear techniques for modeling physiological systems [Chon et al., 1998]. Several neural network control systems have utilized model-based approaches in which the neural network is used to identify a forward nonlinear system... 

A biomedical control system that utilizes a neurophysiologically-based approach has been developed for use in Functional Neuromuscular Stimulation (FNS) systems [Abbas, 1995 Abbas and Chizeck, 1995). FNS is a rehabilitation engineering technique that uses computer-controlled electrical stimuli to activate paralyzed muscle. The task of a control system is to determine appropriate stimulation levels to generate a given movement or posture. The neural network control system utilizes a block diagram structure that is based on hierarchical models of the locomotor control system. It also utilizes a heterogenous network of neurons, some of which are capable of endogenous oscillation. This network has been shown to provide rapid adaptation of the control system parameters [Abbas and Chizeck, 1995 Abbas and Triolo, 1997] and has been shown to exhibit modulation of reflex responses [Abbas, 1995]. 

This chapter presents an overview of the relatively new field of neural network control systems. A variety of techniques are described and some of the advantages and disadvantages of the various techniques are discussed. The techniques described here show great promise for use in biomedical engineering applications in which other control systems techniques are inadequate. Currently, neural network control systems lack the type of theoretical foundation upon which linear control systems are based, but recently... 

Abbas, J.J. and Chizeck, H.J. 1995. Neural network control of functional neuromuscular stimulation... 

Detailed descriptions of the neural network control systems described in this chapter are provided in Neural Network Control Systems edited by Miller, Sutton, and Werbos and The Handbook of Intelligent Control edited by White and Sofge. 

Graupe, D. and Kordylewaki, H., Artificial neural network control of FES in paraplegics for patient responsive ambulation, IEEE Trans. Biomed. Eng. BME-42 699-707,1995. 

Validation of such flap actuation solutions have been performed in wind tunnel tests on a one-seventh downscaled Bell-412 Mach-scaled rotor hub [87]. It has been shown that trailing edge deflections of 4° to 5° can be achieved at up to 1800 rpm which allowed suppression of vibratory bending moments imder an open loop control condition. Even some preliminary closed-loop tests using a neural network controller were performed which however required simultaneous actuation of all four blades. In [88] an induced-shear piezoelectric actuator has been described to actuate trailing edge flaps... 

Hybrid applications of active and passive devices are also known. For example, magneto-ibeological dampers are successfully used as a part of base isolation systems (Ribakov, 2002). Selective control is an effective algorithm for such systems (Ribakov, 2003). A hybrid isolation system, comprised of a bidirectional roller-pendulum system and augmented by controllable magnetorheological dampers is proposed to reduce the potential for damage to stmctures and sensitive equipment (Shook, 2007). Comparison of neural network control, LQR/clipped opti-... 

This error signal is generated at the output of the plant and is passed backward to the neural network controller through the plant and is minimized with the steepest descent method. 

Lieske SP, Thoby-Brisson M, Telgkamp P, Ramirez JM. Reconfiguration of tbe neural network controlling multiple breathing patterns eupnea, sighs and gasps [see comment]. Nat Neurosci 2000 3 600-607. 

Generalized Predictive Control and Bioengineering, by M. Mahfouf and D. A. Linkens Sliding Mode Control theory and applications, by C. Edwards and S. K. Spurgeon Neural Network Control of Robotic Manipulators and Nonlinear Systems, by F. L. Lewis, S. Jagannathan and A. Yesildirek. 

D.M. Katie, and M.K. Vukobratovic, Highly efficient robot dynamics learning by decomposed connectionist feed-forward control, IEEE Trans, on syst. man and cybern., Vol. 25, No. 1, (1995), pp. 145-158. F.L. Lewis and A. Yesildirek, Neural network control of robot manipulators and nonlinear systems, Taylor Francis, (1999). 

---