LAMIH Seminar "Control of discrete time systems under variable input saturation: Application to autonomous freight train".

Summary

Physical, safety, or technological constraints mean that control actuators cannot provide signals of unlimited amplitude or unlimited response speed. Neglecting actuator saturations can be a source of undesirable or even catastrophic behavior for the closed-loop system (such as loss of stability). Several techniques have been developed for the design of control laws for this type of system. Anti-windup is a classical approach to deal with non-linear effects due to input constraints, and override is a related technique to deal with output constraints. This research deals with the synthesis of control algorithms for discrete-time systems under variable input saturation. The control architecture consists of linear switched controllers and a static anti-wake-up compensator (system augmentation).  The problem is formulated using the formalism of switched systems with an arbitrary switching law. Then, sufficient conditions are derived to guarantee the local asymptotic stability and L2 performance of the closed-loop system through a convex optimization procedure formulated by linear matrix inequalities (LMI) where the controller and anti-reversal parameters are computed in a single step. The proposed approach is applied on a real autonomous freight train where the control signal is limited by variable constraints due to adhesion conditions and actuator boundaries.