基于模糊神经网络建模的RFID室内定位算法  

AN RFID INDOOR LOCATION ALGORITHM BASED ON FUZZY NEURAL NETWORK MODEL

作　　者：陈增强[1,2] 国峰[1,2] 张青[2] 

机构地区：[1]南开大学计算机与控制工程学院,天津市智能机器人技术重点实验室,天津300071 [2]中国民航大学理学院,天津300300

出　　处：《系统科学与数学》2014年第12期1438-1450,共13页Journal of Systems Science and Mathematical Sciences

基　　金：国家自然科学基金(61174094);天津市自然科学基金(14JCYBJC18700,13JCYBJC17400)资助课题

摘　　要：将模糊神经网络FNN应用于基于RFID技术的室内定位系统IPS,提出一种基于模糊神经网络的RFID室内定位算法,算法将参考标签数据作为神经网络的训练样本,建立"标签接收信号强度与标签读写器间距离RSSI-DIST"的映射模型,然后利用最小二乘解确定目标的位置坐标.同时,对比了传统BP神经网络和FNN网络在建模和定位中的性能.在仿真和硬件平台测试中,模糊神经网络都要比BP表现出更优异的性能,表明基于模糊神经网络的算法更适合于IPS系统.The fuzzy neural network（FNN） was applied to indoor Positioning System（IPS） base on RFID technology.A RFID indoor Location algorithm based on FNN is proposed,which uses the reference tag data as training samples for neural network and builds the mapping model of ＂The e-Tag received signal strength indicator and the distance between e-Tag and Reader＂（RSSI-DIST）.The least-squares solution is utilized to determine the position of target tag.Meanwhile,the performances in the modeling and positioning were compared between the traditional BP Neural Network and FNN.The Simulation and Hardware platform＇s result prove that FNN shows superior performance than BP and the method based on FNN is more suitable for the IPS.

关 键 词：室内定位系统IPS RFID技术 BP神经网络 模糊神经网络 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]