Iterative learning control

Iterative learning regulations (ILR) are procedures that use the cyclical operation of a process to optimize the control of the process via the cycles. The control error for calculating the setting trajectory is directly influenced by the measurement disturbances. The use of time-variant filters allows the control quality to be improved .

In the operational stability , so-called follow-up tests are carried out in which a recorded load-time signal is to be reproduced as precisely as possible on the test bench. The behavior of the test bench and component means that the actual signal does not match the desired setpoint signal, despite the optimally set controller. In order to improve the tracking behavior (correspondence between the target and actual signal), the target signal is changed in an iterative process until the actual signal matches the original target signal.

The iteration process can be accelerated by first identifying the transmission behavior of the controlled system - i.e. test bench and component . For this purpose, the amplitude and phase response of the controlled system are determined. A so-called drive signal can be generated from the setpoint signal using a mathematical trick, inverting the amplitude and phase response . If you put this drive signal on the test bench, the test bench should reproduce the target signal. However, the actual signal often does not yet quite correspond to the target signal, so that the drive signal has to be changed further in order to obtain the desired target signal.