Due to 5G's stringent and uncertainty traffic requirements,open ecosystem would be one inevitable way to develop 5G.On the other hand,GPP based mobile communication becomes appealing recently attributed to its str...Due to 5G's stringent and uncertainty traffic requirements,open ecosystem would be one inevitable way to develop 5G.On the other hand,GPP based mobile communication becomes appealing recently attributed to its striking advantage in flexibility and re-configurability.In this paper,both the advantages and challenges of GPP platform are detailed analyzed.Furthermore,both GPP based software and hardware architectures for open 5G are presented and the performances of real-time signal processing and power consumption are also evaluated.The evaluation results indicate that turbo and power consumption may be another challengeable problem should be further solved to meet the requirements of realistic deployments.展开更多
文摘针对基于蜂窝信号的室内定位问题,提出基于深度学习的室内定位(Deep Learning-based Cellular Signal Indoor localization,DLCS)算法。DLCS算法建立深度神经网络(Deep Neural Network,DNN)模型,进而学习从信号塔所接收的信号强度(Received Signal Strength,RSS)与用户位置间的非线性关系。DLCS算法由离线阶段和在线阶段构成。在离线阶段,将从信号塔所接收的信号训练学习模型DNN。在在线阶段,用户在监测区域自由移动,从信号塔获取RSS值,再将这些RSS值反馈给DNN,进而估计用户位置。仿真结果表明,提出的DLCS算法能够获取高的室内定位精度。
基金funded in part by National Natural Science Foundation of China(grant NO.61471347)National S&T Mayor Project of the Ministry of S&T of China(grant NO.2016ZX03001020-003)+1 种基金key program for international S&T Cooperation Program of China(grant NO.2014DFA11640)Shanghai Natural Science Foundation(grant NO.16ZR1435100)
文摘Due to 5G's stringent and uncertainty traffic requirements,open ecosystem would be one inevitable way to develop 5G.On the other hand,GPP based mobile communication becomes appealing recently attributed to its striking advantage in flexibility and re-configurability.In this paper,both the advantages and challenges of GPP platform are detailed analyzed.Furthermore,both GPP based software and hardware architectures for open 5G are presented and the performances of real-time signal processing and power consumption are also evaluated.The evaluation results indicate that turbo and power consumption may be another challengeable problem should be further solved to meet the requirements of realistic deployments.
