This study is dedicated to design a fuzzy neural network (FNN) control scheme for performing precision positioning of a particularly-designed four-wire-suspended lens actuator (often called "pickup"), which is capable of moving not only in the conventional translational degrees of freedom (DOFs) - focusing and tracking, but also in the rotational DOF of tilting. This additional tilting DOF provides the capability of the pickup to suppress the non-zero tiling angle of the pickup, which is caused by uneven magnetic fields and other factors of manufacturing/assembling tolerance. The FNN scheme aims to successfully control the aforementioned three-DOFs of the pickup simultaneously for achieving fast focusing, tracking and suppressing non-zero-tilting simultaneously. The design of this FNN control scheme is started with an establishment of a electro-dynamic model of the pickup with considerations of dimensional/assembling errors as parametric uncertainty and unbalanced radial vibrations as disturbance. Three independent FNN schemes are synthesized next to accomplish fast tracking, focusing and zero-tilting simultaneously. Numerical simulations with considering practical system parameters for four different cases with/without the tilting degree and/or uncertainty and disturbance are conducted to verify the effectiveness of the FNN schemes designed. Simulated results show that the designed FNN control scheme is capable of achieving fast zero-tilting with favorable tracking/focusing performances and also robust to unbalanced radial vibrations and assumed plant uncertainty.
- Four-wire-type pickup
- Tilting and fuzzy neural network