Application of polynomial methods for design of adaptive decentralized control

Abstract

The paper deals with the control of multivariable systems with the same number of inputs as the number of outputs using a decentralized approach. The controlled system is divided into single input - single output (SISO) subsystems and adaptive controllers are used to control each subsystem. This paper is focused to the usage of self-tuning controllers (STC) which are a subset of adaptive controllers characterized by on-line identification of controlled system. The control part of individual controllers are based on polynomial approach (dead-beat in the weak and strong version, pole assignment in the 1DOF and 2DOF configurations, controllers based on minimization of quadratic criterion). A supervisory logic block is used in the control scheme to increase the stability of the on-line identification process by ensuring that just one of the controllers is in adaptive mode at a particular time. The logic supervisor is purposed as a simple and reliable technique for designing the supervisory logic. The basic idea of the logic supervisor is based on assigning priorities to individual outputs of controlled multivariable system and on-line evaluation of criterions for each subsystem. The usage of this approach has been tested in simulation experiments and in real-time applications and some results are presented in the paper. © 2004 IEEE.