基于改进BP神经网络的井下无线定位算法研究

Research on Underground Wireless positioning Based on Improved BP Neural Network

针对煤矿井下作业人员安全,提出一种无线定位算法。利用无线网络,采集采样点接收到各个AP的MAC地址和信号强度RSSI,与采样点坐标位置组成训练集;用训练集对加入动量项的BP神经网络进行离线训练,训练完成后,待测点接收到MAC和RSSI信息带到训练过的BP神经网络进行定位。实验结果显示,该定位算法提高了收敛速度和定位精度。

In view of the safety of coal mine workers,a wireless location algorithm is proposed.Using wireless network,the sampling points receive the MAC address and signal strength RSSI of each AP,and form training set with the coordinate position of sampling points.The training set is used to train the BP neural network with momentum term offline.After the training is completed,the measured point receives MAC and RSSI information and takes it to the trained BP neural network for positioning.The experimental results show that the algorithm improves the convergence speed and location accuracy.