Adaptive Filtering and GPC

This paper addresses the question of a suitable "control-relevant identification" strategy for a class of long-range predictive controllers. It is shown that under certain conditions the best process model for predictive control is that which is estimated using an identification objective function similar to the control objective function. The resulting nonlinear least squares calculation is then shown to be asymptotically equal to a standard recursive least squares with an appropriate (model and controller-dependent) FIR data prefilter. Simulation and experimental results demonstrate the validity and practicality of the proposed estimation law.