Title :
A multilayer neural network based identification and control scheme for a class of nonlinear discrete-time systems with asymptotic stability guarantees
Author :
Thumati, Balaje T. ; Jagannathan, S.
Author_Institution :
Dept. of Electr. & Comput. Eng., Missouri Univ. of Sci. & Technol., Rolla, MO, USA
Abstract :
In this paper, a new multi-layer neural network (MNN) based system identification scheme in discrete-time is proposed for a general class of nonlinear discrete-time systems with guaranteed asymptotic convergence of the identification error. Then, a MNN based direct adaptive MNN controller design is introduced for a different class of nonlinear discrete-time systems. The unique aspect of the proposed method is the asymptotic stability assurances of the system identification and tracking errors in the presence of MNN reconstruction errors by using an auxiliary robust term which is a function of the outer-layer NN weights. Finally, simulation examples are presented to illustrate the MNN based estimation and control scheme.
Keywords :
adaptive control; asymptotic stability; control system synthesis; convergence; discrete time systems; identification; neurocontrollers; nonlinear control systems; nonlinear dynamical systems; robust control; tracking; MNN reconstruction error; convergence; direct adaptive MNN controller design; guaranteed asymptotic stability; multilayer neural network; nonlinear discrete-time system; nonlinear dynamical system; robust control; system identification; tracking error; Adaptive control; Asymptotic stability; Control systems; Convergence; Multi-layer neural network; Neural networks; Nonlinear control systems; Programmable control; Robust stability; System identification;
Conference_Titel :
Control and Automation, 2009. MED '09. 17th Mediterranean Conference on
Conference_Location :
Thessaloniki
Print_ISBN :
978-1-4244-4684-1
Electronic_ISBN :
978-1-4244-4685-8
DOI :
10.1109/MED.2009.5164598
