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 Sys- tem (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 indi- cator 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 perfor- mances 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.