Simulation for Improving Error Registration with Neural Network on Base of Back Propagation
Huang Ling-fang · Jisuanji fangzhen · 2010
The widely used BP network models do not guarantee the convergence to the? global? minimum? point.? To the eliminate network errors,improve the convergence rate is proposed in BP networks by adding feedback signals generated within the recurrent neural network error matching algorithm.Algorithm in the internal recurrent neural network introduces the last time output by adding a priori knowledge and to improve the convergence speed.At the same time,,the back - propagation learning rule for deviation unit recurrent neural networks was derived,with the network the cumulative error less or equal to the required value.The simulation of civil aviation radar network system simulation data shows that the algorithm which eliminates radar network system error and improves target precision can obtain good results.