【摘要】 基于固定参数的无线信号传播损耗模型的定位算法,不能很好解决由于多径传播效应和环境复杂性所带来的测距误差问题。提出使用GRNN神经网络来拟合室内RSSI值与距离值之间的映射关系,得到RSSI值与距离值的映射模型,再将定位实验中实测的RSSI值作为训练好的GRNN神经网络的输入层,在输出层得到与RSSI值相对应的距离值,最后使用加权质心算法来进行待测节点的定位。该算法不仅简单而且性能良好,并且不需要额外的硬件。经过Matlab和ZigBee实验仿真验证,与路径损耗模型和基于BP神经网络的定位算法相比,所提出的算法可以提供较好的定位结果。更多还原

【Abstract】 In response to the problem that the localization algorithm based on the wireless signal propagation loss model with fixed parameters can’t remove ranging errors induced by multipath propagation effects and environmental complexity. This study adopted GRNN neural network to fit the RSSI value and distance value, and then get the mapping model of RSSI value and distance value. It adopted the RSSI value as the input layer of the trained GRNN neural network and derived the RSSI value in the output layer. Finally,the weighted centroid algorithm was applied to locate the node. It finds that the algorithm is simple and well-behaved,and does not require additional hardware. Through the simulation and experiment results on the MATLAB and ZigBee, compared with the localization algorithm based on the path loss model and BP neural network, the proposed algorithm can provide better localization results.更多还原