FPGA implementation of a multilayer perceptron neural network using VHDL

Our aim is how to use the run-time reconfiguration ability of SRAM-based FPGA circuits in older to implement artificial multilayer neural networks which are needed to process, in real-time, nonlinear data obtained from a set of microchemical sensors. So we propose, in this paper, an innovative alternative to ASICs and/or multiprocessors based architectures.