Hardware implementation of Radial Basis Function Neural Network based on sigma-delta modulation

A digital hardware implementation of Radial Basis Function Neural Network (RBFNN) based on sigma-delta modulated bit-streams is presented. Through the change of feedback coefficient in the framework of traditional sigma-delta modulator, a new limiting amplifier modulator (LAM) is fabricated, and the approximation to Gauss kernel function can be achieved by the combination of several LAMs with different coefficients. The bit-stream neurons with Gauss kernel function and a whole feed-forward artificial neural network is implemented on field programmable gate array (FPGA). Thus a nonlinear function approximation problem can be solved by the neural networks presented.