Repetitive and adaptive feedforward schemes for high performance control of automatic machines mechanisms

In this thesis different high performance controls of non-linear mechanisms are presented and evaluated. The key idea of the proposed schemes consists in modifying the periodic reference trajectory provided to the plant in order to reduce as much as possible the tracking error by compensating unmodelled dynamics and periodic disturbances that cyclically affect it. Starting from the study of the standard Repetitive Control structure, the purpose was to design a new and more flexible control algorithm that could avoid the several problems presented by the standard RC technique. According to this aim two modified versions of the repetitive control and an adaptive feedforward approach are developed and their simulation results are discussed. Moreover, this work was done in collaboration with the Line Control & Technology office of the G.D S.p.A. company and the experimental results obtained with an industrial setup composed by a Beckhoff PLC, a Bosch drive and Kollmorgen motor, are reported only for the adaptive feedforward approach, showing the real effectiveness of the proposed method.