BP多传感器融合技术在罐车油位误差处理中的应用  

The application of BP multi-sensor fusion in tank oil level's error processing

作　　者：杨洪军[1] 王振友[1] 程海涛[2] 

机构地区：[1]山东理工大学计算机科学与技术学院,山东淄博255091 [2]青岛理工大学费县校区,山东临沂273400

出　　处：《山东理工大学学报（自然科学版）》2011年第3期28-30,35,共4页Journal of Shandong University of Technology:Natural Science Edition

摘　　要：在油罐车油位测量中,超声波传感器受本身结构和外界温度等因素的干扰,其输入输出特性呈明显的非线性,仅靠硬件补偿的方法来提高精度,效果不明显.采用BP多传感器融合来实现多超声波传感器的非线性误差校正,大大提高了超声波测距精度.另外,BP网络模型很好的泛化能力,简单的网络结构,简明的算法,使其在测控系统及智能传感器中具有良好的应用前景和推广价值.In the measurement of oil level in the tanker,the input-output linearity of the ultrasonic sensor is not always a straight line,mainly interfered by the structure,outside temperature and other factors.Thereby it will not be improved very much merely by the method of hardware compensation.In this paper,BP multi-sensor fusion is used to realize non-linear error correction of ultrasonic sensor,and ultrasonic distance measurement accuracy was greatly improved.In addition,BP network model has a good generalization ability,a simple network structure and simple algorithm,so that it has good prospects in the control system and intelligent sensor.

关 键 词：油位计量 多超声波 BP网络 误差修正 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]