Research on intellectual signal classifier system

In this paper, combined-algorithm classifiers are proposed, which include neural network based on genetic algorithm, constant-coefficient fuzzy neural network, and fuzzy system based on genetic algorithm. `Double helix´ problem is used to compare their performances. These classifiers are applied to distinguish NEMP from LEMP. Due to its excellent classification performance, fuzzy classifier based on genetic algorithm has been chosen to hardware implementation. Based on SOPC (system on programmable chip), with a single large-capacity FPGA as the core controller, an intellectual signal classifier platform is built by the reusable IP technology. The correct classification rate is 100% for NEMP and 92% for LEMP respectively.