神经网络数据融合在位移传感系统中的应用 - 工程技术版.pdfVIP

6 1 () Vo.l 6 N o. 1 2006 3 JOURNAL OF NAN JING NORMAL UN IVER SITY ( ENGINEERING AND TECHNOLOGY) M ar, 2006 神经网络数据融合在位移传感系统中的应用 朱海梅 (, 225002) [] . , . RBF() , , RBF,. : , 4. . [] , BP , RBF , [] T P2126 [] A [] 1672-1292( 2006) 0 1-0028-05 Application of Data Fusion Based on NeuralNetwork for Displacem ent Sensing System ZHU Ha mi e i ( College of Phy ical S cience and Techn ology, Yangzhou U n iver ity, Yangzhou 225002, Ch ina) Abstract: The tem perature dr ift error i one of the m ajor error of the di placem ent en or y tem. In order to mi - prove the accuracy and tability, the output from the di placem ent en or and the one from the temperature en or can be fu ed. A t pre ent, neu ral network ha been w idely app lied to the area of data fu ion, in w hich data from mu ltiple en or arem o tly fu ed by BP neural newt ork. The low training peed of BP newt ork lead to the poor practicab ility, therefore in thi paper, a data fu ion m ethod ba ed on RBF neural network i propo ed for reduc ing temperature drift error. The output from the di p lacem ent en or and tho e from the tem perature en or are ent to the fu ion center, w here the RBF neural newt ork i trained and a a re u lt the table outpu t i obtained. Exper mi ent how that the out- put tab ility of the di p lacem ent en or i mi proved by 4 tmi e under the am e tem perature fluctuation. It i obv iou that the propo ed method in thi paper i effective for reduc ing the drift error of di placem ent en or y tem. K ey words: data fu ion, BP neural network, RBF neural network, temperature dr ift 0 引言 . , , , . , , . , , , . : ,