The compute core of a FPGA is a complicated programmable logic integrated circuit array which different from the ordinal instruction execution of the traditional computer. It changes the compute pattern of traditional compute and provides a new method to realize the high speed compute. Hardware Implementation is very important when considering computational velocity of neural networks (NNs), especially NNs with learning ability implemented by integrated hardware. Firstly, this paper presents the design of FPGA based coprocessor which is the hardware platform well-suited for the implementation of NNs. Secondly, it expounds hardware implementation of RBF (Radial Basis Function) Neural Network, and analyzes the performance and problem of the system. According the data of experiments, the compute speed of the RBF neural network implemented by hardware that realized by FPGA coprocessor is much higher than the speed of compute run on the PC. © Springer-Verlag 2010.
CITATION STYLE
Yang, Z. G., & Qian, J. L. (2010). Hardware implementation of RBF neural network on FPGA coprocessor. In Communications in Computer and Information Science (Vol. 105 CCIS, pp. 415–422). https://doi.org/10.1007/978-3-642-16336-4_55
Mendeley helps you to discover research relevant for your work.