On dead-time competation in repetitive control
Repetitive control is an elegant implementation of the internal model principle, which incorporates a model of an arbitrary periodic signal via the use of a delay element. The main difficulty in the design of controllers incorporating repetitive elements is the need to analyze the stability of an unorthodox time-delay system.
Existing solutions often do that via robust control arguments, leading to conservative solutions. In this talk I put forward an alternative approach, which reduces the stabilization problem to that for a delay-free system.
The proposed architecture is, in a sense, dual to that used in various dead-time compensation schemes, like the Smith predictor. In the minimum-phase case, an implementation scheme, which is insensitive to the value of the delay in the repetitive block, is proposed.