Constrained Controller and Observer Design by Inverse Optimality

Model predictive control (MPC) is often tuned by trial and error. When a baseline linear controller exists that already well in the absence of constraints MPC introduced to enforce them, one would like avoid altering original feedback law whenever they are not active. We formulate this problem as matching similar works Di Cairano Bemporad (2009), (2010), Tran et al. (2015), which we extend more general framework. prove positive-definite stage-cost matrix yielding property can be computed for all stabilizing controllers. In addition, constrained estimation also solved similarly, observer with moving horizon estimator. Finally, discuss various aspects practical implementation proposed technique some examples.