Improved repetitive control scheme and its application to a 1DOF lagrangian system

The aim of this thesis work is the developement of a new and more flexible implementation scheme of the well-known repetitive control (RC). Different innovative repetitive control schemes have been implemented in Simulink (introducing some signal elaboration such as a linear interpolation) and then simulated on a simple “pure delay” discrete-time system with no synchronization, and on a more complex and realistic 1DOF lagrangian system representing the four-bar linkage of the ACMEC project on which the repetitive control will be tested in practice. A comparison between the standard schemes and the developed one is carried out and the differences are highlighted. In particular it is shown that the new implementations allows to work significantly better (lower residual error and faster transient) with respect to the classical schemes, in case of error period not multiple of the sampling time of the control. More than that they are easier to implement on digital platforms since the memory required is chosen independently from the error period (so from the machine speed) and from the sampling frequency of the control