Data-Driven Control of Linear Direct-Drive Motors


To improve the tracking performance of linear motors several data-driven approaches to tune the pre-compensator parameters of a two-degree of freedom controller are proposed. These approaches are applied on a two-axis direct-drive high precision positioning set-up. Three different types of reference signals are considered:

1. The desired trajectory is generated by a reference model. The tuning objective is to minimize a norm of the error between the achieved closed-loop system and the reference model. The IFT, VRFT and ICbT approaches are used to reduce the two norm of the output error or to make the output error uncorrelated with the reference signal.

2. The reference signal is periodic. in this case the Iterative Learning Control algorithm is used to improve the tracking performance of the system. It will be tried to make the algorithm insensitive to stochastic perturbations.

3. The reference signal is perfectly known. In this case, the methods based on preview optimal tracking will be employed.