Feed forward adaptive control of a linear brushless DC motor

Design and fabrication of a learning controller of a brushless DC linear motor is investigated in this paper. The controller includes two main parts termed as feedback and learning adaptive feed forward controller. The feedback is preformed based on a PID method by linear Second order model of a linear motor. The learning adaptive feed forward controller is based on neural network with one hidden layer second-order B-sptine basis functions and is used for prediction the un-modeled forces and unknown relations. By analyzing the effective forces in movement, simple model of motor can attain. However prediction of a full model of motor can be too complex. Therefore by using a simple model of the motor, the exact control can be achieved by a feedforward controller. Simulation and real test are shown the proper performance of the suggested method due to low effort and complexity.