SLIDING OBSERVER-BASED NONLINEAR MODEL-PREDICTIVE CONTROL

State observer based nonlinear model predictive control approaches hav e been reported in literature to control open-loop unstable nonlinear processes, The inherent drawback of these linearization-based techniqu es is the requirement of accurate state estimates to ensure the qualit y of the linearized model, In this article, a rigorous nonlinear state estimator based on sliding mode is thus proposed and is incorporated with the nonlinear quadratic dynamic matrix control (NLQDMC) for the r efinement of implementation and performance, The unknown disturbances or modeling errors existing in a nonlinear system can be modelled and estimated as augmented states by using this sliding observer. Two exam ple problems are utilized to demonstrate the construction and performa nce of this sliding observer based nonlinear predictive control for op en-loop stable and unstable processes.