| 摘要: |
| 为了提高远距离无线电（Long Range Radio，LoRa）系统定位精度，提出利用接收信号强度指示（Received Signal Strength Indication，RSSI）测距和三边定位方法结合小波神经网络模型对LoRa节点进行定位。首先分析了RSSI测距、三边定位的原理和实现方法，给出了基于RSSI测距方法求解LoRa网关与LoRa节点间距离的步骤。然后利用3层融合型小波神经网络搭建了LoRa节点定位模型，选取Morlet小波为隐含层神经元的激励函数，选取Sigmoid阈值函数为输出层函数，将LoRa节点到3个LoRa网关的距离作为输入层数据，节点定位模型转换并输出LoRa节点位置的归一化横坐标和纵坐标。最后利用LoRa网关模块和LoRa节点模块搭建了LoRa节点定位实验系统并进行了实验测试，通过实验数据分析得出了RSSI测距参数并利用三边定位算法和小波神经网络节点定位模型实现了待定位LoRa节点的精准定位。实验结果表明，所提算法的定位精度为1.033 m，优于四点质心定位算法和传统三边定位算法。 |
| 关键词: LoRa定位 RSSI测距算法 三边定位算法 小波神经网络 |
| DOI：10.20079/j.issn.1001-893x.240924001 |
|
| 基金项目:2023年广东省科技创新战略专项资金项目（pdjh2023b0844）；2024年广东省高等教育“冲补强”特色高校提升计划项目（31500022） |
|
| A LoRa Positioning System Combining RSSI Ranging and Trilateration Algorithm |
| SHI Zhenjiang,LI Junyi,HUANG Jiajia |
| (School of Engineering and Technology,Open University of GuangdongGuangdong Polytechnic Institute,Guangzhou 510091,China) |
| Abstract: |
| To improve the positioning accuracy of long range radio(LoRa) systems,received signal strength indication(RSSI) ranging and trilateration methods combined with wavelet neural network model are proposed to locate LoRa nodes.Firstly,the principles and implementation methods of RSSI ranging and trilateration are analyzed,and the steps are provided for solving the distance between LoRa gateway and LoRa node based on RSSI ranging method.Then,a 3-layer fusion wavelet neural network is used to build a LoRa node localization model.Morlet wavelet is selected as the excitation function for the hidden layer neurons,and Sigmoid threshold function is selected as the output layer function.With the distances from LoRa nodes to three LoRa gateways as input layer data,node localization model converts it and output the normalized horizontal and vertical coordinates of LoRa node positions.Finally,a LoRa node positioning experimental system is built using the LoRa gateway module and LoRa node module,and experimental tests are carried out.The parameters in the RSSI ranging mathematical model are obtained through experimental data analysis,and the precise positioning of the LoRa node to be located is achieved using the trilateration algorithm and wavelet neural network node localization model.The experimental results show that the proposed algorithm has a positioning accuracy of 1.033 meters,which is superior to that of the four point centroid positioning algorithm and the traditional trilateration algorithm. |
| Key words: LoRa localization RSSI ranging algorithm trilateration algorithm wavelet neural network |
