New rejection methods of control system based on Forward Propagation neural networks
前向FP神经网络实现控制系统的拒识模式

It is very important to deal with the relation between incorrect recognition and rejection in a control system because the former will do much more harm than the latter.A new rejection method based on Forward Propagation (FP) neural networks is proposed.Through the study of sample-sets of instruction and state-information codes of control system,the FP neural networks can not only ensure that every code has the biggest attract radius,namely the control system has the greatest correction ability and fault tolerance,it also gives the 100% rejection to the signals outside the attraction radius.So it can increase the rate of correction recognition to ensure the system to run safely and smoothly.