Research on the Hardware RBF Fuzzy Neural based on the FPGA

To meet the real-time requirements of industrial field, the hardware FPGA RBF fuzzy neural network is designed and implemented based on the 250000 door Spartan-3E (XC3S250E) chip of Xilinx First the structure and algorithm of RBF fuzzy neural network is introduced, and then a kind of improved hybrid excitation function approximation algorithm is put forward in order to overcome the difficulties in the hardware design of the neural network, and the hardware co-imitation and timing simulation is done on it. The experimental results demonstrates that, the method has better identification precision and speed, and it is a effective method of hardware implementation of RBF fuzzy neural network, and it lays the foundation for control and image processing based on hardware neural network.