基于阵列神经网络技术的多传感器信息融合  

Information Fusion with Multisensor Based on Array Neural Network Technique

作　　者：李华星[1] 高社生[1] 杨波[1] 

机构地区：[1]西北工业大学,西安710072

出　　处：《弹箭与制导学报》2004年第S6期366-368,共3页Journal of Projectiles,Rockets,Missiles and Guidance

摘　　要：将阵列神经网络技术应用于多传感器信息融合系统中,利用阵列神经网络分布式信息存储和并行处理数据的特点进行信息处理,减少了由于模型不符合和特征选择不当带来的误差,提高了多传感器系统的性能.提出了一种基于阵列神经罔络滤波器的 INS/GPS/SAR 组合导航信息融合方法。仿真结果表明:基于阵列神经同络滤波器的 INS/GPS/SAR 组合系统的导航精度要比传统的卡尔曼滤波器的精度要高。The array neural network technique is applied to muhisensor information fusion system.With the distributed information storing and the parallel data processing technique introduced to the information processing,the errors coming from the mismatched model and improper feature selection are decreased,meanwhile,the performance of the muhisensor system is improved.A method of array neural network filter is presented for INS/GPS/SAR integrated navigation system in this paper.The simulated result indicated:The precision of the INS/GPS/SAR integrated naviga- tion system based on the array neural network filter is much higher than that of the navigation system of traditional Kal- man filter.

关 键 词：阵列神经网络 信息融合 组合导航系统 

分 类 号：TP202[自动化与计算机技术—检测技术与自动化装置]