An improved particle swarm optimization-based dynamic recurrent neural network for identifying and controlling ultrasonic motors

A learning algorithm for dynamic recurrent Elman neural networks is proposed based on an improved particle swarm optimization. The proposed algorithm performs the evolution of network structure, weights, initial inputs of the context units and self-feedback coefficient of the modified Elman network together. A novel control method is presented successively based on the proposed algorithm. A novel dynamic identifier is constructed to perform speed identification and also a controller is designed to perform speed control for ultrasonic motors. Numerical results show that the designed identifier and controller based on the proposed algorithm can both achieve higher convergence precision and speed. The identifier can approximate the nonlinear input-output mapping of the USM quite well, and the good control effectiveness of the controller is verified using different kinds of speeds of constant, step, and sinusoidal types. Besides, the preliminary examination on the randomly perturbation also shows the fairly robust characteristics of the two models.