The systematic design of control laws for nonlinear systems subject to state and input constraints is one of the major challenges in the control of real-world systems. Depending on the application, the desired solution must achieve a satisfactory trade-off between multiple and often contradicting requirements such as optimality, computational efficiency, robustness, reliability, tuning simplicity, and generality. This article proposes a novel framework for the closed-form control of nonlinear systems subject to constraints on the state and input variables.In the last two decades, Model Predictive Control (MPC) has established a golden standard for constrained control problems. In MPC schemes, the control input is generated at each time step by solving an online optimization problem on the predicted state trajectories [1]- [5]. A strong point of MPC schemes is that they not only solve the constrained control problem, but they may also optimize some performance index. As such, MPC typically outperforms other solutions in terms of output response. Several MPC schemes have been proposed in the literature for different class of systems and constraints, and have been successfully applied in various realworld applications [6], [7]. However, MPC also presents some important drawbacks, the main of which is that, being based on online optimization, it has an elevated computational complexity. As such, one of the main objectives of current MPC research is the development of increasingly efficient solvers, especially for nonlinear problems [8]-[10].An alternative optimization-based strategy for constrained control are Reference Governor (RG) approaches. RGs can be seen as special MPC schemes that are specifically designed for prestabilized systems [11], [12]. Due to the particular properties of the resulting online optimization problem, RG schemes are typically computationally less onerous than other MPC strategies, although they also tend to provide reduced performances [13]- [15]. Nevertheless, since they are optimization-based approaches, RG solutions can still be problematic for those real-world systems that have to comply with strict real-time requirements and/or where the use of powerful control units is not economically realistic. This fact advocates for the development of simpler
