摘要
提出一种新的带有神经网络的GPS/DR组合数据融合算法以求在动态路径诱导中获得实时定位数据。数据融合算法采用的是具有单隐层的 3层神经网络 ,并给出了里程仪、速率陀螺和DR系统的数学模型以及里程仪的标定因子的修正模型。它能有效克服因丢失卫星信号而造成定位精度下降的现象。
A new data fusion algorithm with neural network in GPS/DR integrated positioning system is presented in order to get precise positioning data. The mathematical model of odometer, gyroscope and DR system is given. A three-layer neural network that has single hidden layer is adopted in the framework of data fusion algorithm. According to the experience equation eight hidden layer neurons is suitable. Moreover, the emendatory method of odometer demarcated gene also is given. Therefore, it can effectively maintain the positioning precision not to be lowered owing to the loss of satellite signal.
出处
《公路交通科技》
CAS
CSCD
北大核心
2000年第z1期67-70,共4页
Journal of Highway and Transportation Research and Development
基金
国家自然科学基金资助项目! ( 596382 2 0 )
