基于深度学习与信息交互的5G终端群组定位方法

Research on 5G Terminal Group Positioning Method Based on Deep Learning and Information Exchange

针对传统无线定位问题中基于接收信号强度指示(Received Signal Strength Indicator,RSSI)的测距与定位方法精度过于依赖RSSI测距精度,以至于提升定位精度成本较大的问题,提出了基于深度学习与信息交互的第5代移动通信技术(5G)终端群组定位方法,以降低其对基站与终端间RSSI测距精度的依赖。在5G终端群组条件下,基于终端间的相互测距信息,利用终端间测距误差较小的特点来弥补基站与终端间RSSI测距的误差,并结合深度神经网络,将接收到的RSSI信号作为输入特征、位置信息作为输出特征,进行模型训练并输出5G终端群组定位结果,使得最终定位精度得到有效提升。仿真试验验证了所提出方法的有效性。

To address the problem that the accuracy of the ranging and positioning method based on RSSI(Received Signal Strength Indicator)in the conventional wireless positioning technology is seriously dependent on the RSSI ranging accuracy,which results in a high cost of improving the positioning accuracy,this paper proposes a novel 5G terminals group positioning method based on deep learning and information exchange,in order to reduce its dependence on the RSSI ranging accuracy between base stations and 5G terminals.Under the condition of 5G terminals group,based on the mutual ranging information between 5G terminals,the smaller ranging error between 5G terminals is utilized to compensate for the RSSI ranging error between base stations and 5G terminals.Combined with DNN(Deep Neural Networks),the received RSSI signals are used as input features,and the positioning information is used as output features to train the model and output the 5G terminals group positioning results,so that the final positioning accuracy is effctively improved.Simulation experiments verify the effectiveness of the proposed method.