Model predictive control is one of the promising advanced process control techniques for industries. The control algorithm is based on modelling of the system/s and an optimization process to generate the control action. State space models are commonly used in the nonlinear model predictive control. The control action is optimized basing on the knowledge of current state information . Besides, state and output feedbacks are often required to design a stabilizing control performance while satisfying the control and state constraints . However, some states can’t be measured directly. A state observer (filter) is therefore used to estimate the states in the current iteration such as Kalman filter or Luenberger observer[3, 4]. However, sudden disturbances or faults will introduce large estimation error from the observer; the computed control action based on the faulty estimated state will be error prone. To deal with the sudden disturbance or faults, an unknown input observer is proposed to estimate the states. The unknown input observer aims to deal with the disturbance, which is assumed to be larger than a regular measurement/process noise (white noise). In this paper, a nonlinear unknown input observer is designed based on a nonlinear process model. The state estimation from unknown input observer was applied for the optimization process. An MPC stability theorem dealing with the asymptotic stability of the observer was proposed. An example of multi-variable batch reactor was used to demonstrate the proposed approach.
 Tang G-Y, Guo Z-W. Model predictive control based on state predictive observers for discrete systems with time-delay. Control and Automation, 2003 ICCA'03 Proceedings 4th International Conference on: IEEE; 2003. p. 365-9.
 Findeisen R, Imsland L, Allgower F, Foss BA. State and output feedback nonlinear model predictive control: An overview. European journal of control. 2003;9:190-206.
 Prakash J, Senthil R. Design of observer based nonlinear model predictive controller for a continuous stirred tank reactor. Journal of Process Control. 2008;18:504-14.
 Qian J, Dufour P, Nadri M. Observer and model predictive control for on-line parameter identification in nonlinear systems. IFAC International Symposium on Dynamics and Control of Process Systems (DYCOPS)2013. p. 571-6.