Feedback-error-learning neural network for trajectory control of a flexible micro-actuator

Summary form only given. A neural network approach to online learning control and real-time implementation for a flexible microactuator is presented. The flexible microactuator is made of a bimorph piezoelectric high-polymer material (poly vinylidene fluoride). The bimorph piezoelectric microactuator consists of two PVDF films cemented with a metal shim in proper polarity. When a voltage is applied across the elements, the upper layer elongates, while the lower contracts producing a deflection. The control scheme consists of a feedforward neural network controller and a fixed-gain PID controller. This neural network controller is trained so as to make the output of the PID controller zero. In the process, the neural network learns the inverse dynamics of the system. The feedback-error-learning controller is implemented using a digital signal processor, and experimental results are given.