基于改进指纹地图定位算法的LoRa定位方法

LoRa localization method based on improved fingerprint map localization algorithm

针对LoRa网络节点定位过程中,接收信号强度指示(RSSI)采样值波动性大,指纹地图定位不精确的问题进行研究,提出了一种改进的指纹地图定位算法,即PSO-ANN-FMLA。首先,通过人工神经网络算法获得最佳RSSI值,利用粒子群算法优化人工神经网络算法的隐含层的神经元数量和学习速率;然后,指纹地图定位算法选择最佳RSSI采样值进行定位。最后,通过实验验证了本文算法的有效性,在室外300 m×300 m的无遮挡区域内,平均定位精度为6.21 m;与传统指纹地图定位算法相比,本文提出的算法平均定位精度提高了59.86%。

Aiming at the problem that large fluctuation of received signal strength indication(RSSI)sampling value sampled in accurate fingerprint map localization in LoRa process of network node localization.An improved fingerprint map localization algorithm is proposed, Firstly, the optimal RSSI values is obtained by artificial neural network(ANN) algorithm, in which the particle swarm algorithm optimizes the number of hidden layer neurons and learning rate of ANN algorithm, then fingerprint map localization algorithm uses the best RSSI value to localize.Finally, the effectiveness of this method is verified by experiment.The experimental results show that the positioning precision of the proposed algorithm is 6.21 m in the outdoor 300 m×300 m environment without occlusion.Compared with the traditional fingerprint map localization algorithm, the average positioning precision of the improved fingerprint map localization algorithm is 59.86 % higher.