Command governors for constrained nonlinear systems: direct nonlinear vs. linearization-based strategies

Two command governors (CG) are described for solving set-point tracking problems for nonlinear systems when pointwise-in-time input and/or state related inequality constraints have to be fulfilled. A CG is a nonlinear device which is added to a primal compensated system. The CG action, based on the current state, set-point and prescribed constraints, is finalized to select, at any time, a command sequence under which the constraints are possibly fulfilled with acceptable tracking performance. The two CGs, both based on the receding horizon control philosophy, differ in that one directly exploits the nonlinear model of the primal system, whereas the other is based on the property that, under some conditions, all trajectories of a nonlinear system are embedded within the trajectories of a linear time-varying (LTV) uncertain system. An example is presented so as to exhibit the relative merits of the methods.