神经网络和RFID相融合的室内定位算法

Indoor Localization Algorithm Based on BP Neural Network and RFID

在基于RFID的室内定位算法中,路径损耗系数固定,不能反映室内环境复杂多变特性,导致室内定位误差大,为了降低室内定位的误差,提出一种神经网络和RFID相融合的室内定位算法。首先计算室室内的接收信号强度,然后将接收信号强度作为神经网络的输入向量,路径损耗系数作为神经网络的目标输出,通过神经网络的训练建立室内场强信号传播模型,并采用遗传算法优化神经网络的参数;最后采用具体实例对室内定位性能进行测试与分析。结果表明,相对于RFID的室内定位算法,本文算法可以动态、准确估计信号的路径损耗系数,实时反映室内环境的信号变化,提高了室内定位精度,可以有效地满足室内的无线定位要求。

In RFID indoor localization algorithm,path loss coefficient is fixed,but can not reflect the indoor complex and changeable characteristics for environment to cause indoor localization error,in order to reduce the error of the indoor localization,this paper presented an indoor localization algorithm based on neural network and RFID. Firstly,the received signal strength of indoor is calculated,and then the received signal strength is taken as the input vector of the neural network while the path loss coefficient is taken as the target output of neural network,and indoor path loss coefficients model is built by using neural network and the genetic algorithm is used to optimize neural network parameters.Finally,some examples are used to test and analysis the indoor localization performance. The results show that,compared with the indoor localization algorithm based on RFID,the proposed algorithm can accurately estimate the path loss coefficient,and reflect signal changes in the indoor environment,improve indoor localization precision,can effectively meet the localization requirements of indoor wireless.