STATE ESTIMATION AND CONTROL OF NONLINEAR PROCESS USING NEURAL NETWORKS
Anna Jadlovská
This paper considers the use of neural networks for nonlinear state estimation,
identification and control of nonlinear processes. The nonlinear identification
is using feedforward neural networks as useful mathematical tool to build model
between the input and the output of a nonlinear process. In this paper is
considered the possibility an online state estimation of the actual parameters
from offline trained neural state space model of the nonlinear
process using the gain matrix. This linearization technique is used in the
algorithm online tuning of the controller parameters based on pole placement
control design for nonlinear process.
Keywords: dynamic neural models, nonlinear state estimation, gain matrix, nonlinear control
