Nonlinear systems with both state and input constraints are difficult to control. The complexity of
such system is the problem of simultaneous satisfying the nominal control object and each of the
individual constraints. In this report, a method is proposed to design state-feedback control laws that
guarantee state-constraint satisfaction while the control input remains bounded. The main idea is to
use a recursive procedure to construct a constraint feasible set that is positive invariant with bounded
control. Constraint satisfaction is then achieved bymodifying the control input on the boundaries such
that the solution of the system remains inside the feasible set for all time. The proposed method will
be illustrated on two simple design examples at the end of this report.
Traineeship report. - DC 2010.028