Effect of measurement noise on a gradient based repetitive control algorithm

Some theoretical properties, including robustness against multiplicative modelling uncertainty, of a recently developed repetitive control algorithm are reviewed in this paper. In addition new results are presented that analyse, using stochastic analysis, the effect of measurement noise on the algorithm performance. Simulation examples are used to illustrate these new theoretical findings. The algorithm is also applied in real-time in the context of active vibration isolation using a naval mount system. The experimental results demonstrate that the algorithm is capable of reducing a periodic disturbance by 50 dB.