Résumé
In this paper, the track following problem of the Read/Write (R/W) head of a Hard-Disc-Drive (HDD) is addressed using Robust Integral of Sign Error (RISE) based Neural Network (NN) technique. The proposed control scheme is required to compensate as much as possible the nonlinear hysteresis friction behavior which degrades the HDD performance through generating important residual tracking errors. It is well shown that the RISE technique, along with the NN based feedforward control, is able to guarantee the stability of such a system. Moreover, the boundedness of the closed-loop signals is ensured. To the best authors' knowledge, the suggested control solution, applied at the low frequency region of a HDD, has never been conducted before on such system. Different simulation scenarios are performed including nominal case and external disturbance rejection to demonstrate that the proposed solution is robust and efficient to achieve good tracking performances.