|
Abstracts |
V.F. Gubarev, L.M. Yakovleva, N.N. Aksenov, A.V. Palagin, V.N. Opanasenko, V.G. Sakharin
Recently, interest has grown for application of reconfigurable devices in robust and adaptive control systems. The main advantage of such devices is that its structure is not fixed and may be varied depending on the currently used control algorithm. In this paper a new PC-based controllers-reconfigurable microcontrollers (RMC) are offered for experimental physics control systems. Programmable logic devices (PLD) Xilinx Corporation of a type FPGA [1] are used in RMC. Reconfigurable devices are connected to PC as the coprocessor through the system bus (for example, PCI) forming the system with programmable architecture as a result. Each control algorithm is realized as a file of a PLD-configuration. The set of model- based control algorithms make up the functional library. Model selection procedure on the base of control relevant identification is a part of unified control system design [2]. Methodology of the joint and iterative identification and control is proposed which well suited for the realization in RMC and convenient for users. Control design is arranged as follows. First the identification experiment is carried out and approximate model is reconstructed using special identification algorithm and model structure in normalized form. Then select from the library the appropriate liner or nonlinear parametrized control law. After that the parametric class of controllers which guarantee stability of the closed loop system is automatically determined [3]. The best robust controller inside given class obtained such way is evaluated by the parameter tuning procedure, which give opportunity to impact performance of the transient processes in closed system without loosing asymptotic stability.
[1] The Programmable Logic Date Book, Xilinx Inc. 2000
[2] Yakovlev O.S., Method of Structural Synthesis of Nonlinear Regulators, Journal of Automation and Information Sciences, Vol. 30 No. 1, 1998, pp.70-81.
[3] Gubarev V.F. Control Relevant Identification of Distributed Parameter Evolutionary Systems. Proceedings of International Conference SICPRO2000, Moscow, 2000, 547-555 pp.
{*}Institute of Cybernetics, Kiev, Ukraine
{**} Institute of Cybernetics, Kiev, Ukraine
{***} Institute of Cybernetics, Kiev, Ukraine
eCONF
C011127
ePROCEEDINGS
COMMITTEES
PARTICIPANTS
PROGRAM
SUMMARY
ICALEPCS 2001
eCONF
|
||
|
|
Last updated: September 13, 2002 Website Problem? Contact Technical Publications |
 |