The research of electromotor control based on optimized RBF neural network

Ni-qin Jing, Lin-na Wang

COMPUTER MODELLING & NEW TECHNOLOGIES 2014 18(5) 14-18

Beijing polytechnic, Beijing 100015, China

RBF neural network suits to control electromotors, which have uncertainty and highly nonlinear systems. However, in practice, RBF neural network also have some obvious defects. For example, the strong dependence on the initial parameter and the poor quality of clustering algorithm. For the above defects, this paper is going to build an optimized RBF neural network through the combination of ant colony optimization algorithms, chaos ergodicity optimization theory and traditional K-means algorithm. On this basis, the optimized RBF neural network will be applied to PID control and then the dynamic performance of the electromotor will be simulationally tested by the designed PID controller. The simulation results show that in the control of electromotor, the optimized RBF neutral network has the characteristic of high control accuracy and strong traceability and also it has the ability to guarantee electromotor control system with steady and dynamic performance.