BP-Neural-Network-Based Adaptive Control for Trajectory Tracking of a Two-DOF Compliant Micropositioning Platform

The control of compliant mechanisms is mostly implemented using the proportional-integral-derivative (PID) controller, which requires manual adjustment of control parameters and has limited control accuracy. To achieve adaptive control for trajectory tracking on compliant micropositioning platforms, this article proposes an intelligent control method based on the back propagation neural network (BPNN). First, a two-degrees-of-freedom platform based on amplifying mechanisms is introduced, detailing its configuration, prototype manufacturing, and the setup of the experimental system. Then, according to specific control requirements, the structure of the BPNN is determined, integrating it with a PID controller to design a BPNN PID (BPNNPID) controller. To prove its superiority, the control experiments of single axis sine wave trajectory and two-axis elliptical trajectory are carried out by using the traditional PID controller and the BPNNPID controller. The results show that the traditional PID controller exhibits obvious phase delay, while the BPNNPID controller possesses adaptability and it can automatically determine the optimal control parameters in real time. This effectively compensates for phase lag and improves system accuracy. The method realizes the independent optimization and adjustment of control parameters at different frequencies, and it provides a certain reference for the intelligent control of such mechanisms.