基于资源调度和深度学习的5G多用户定位

由于5G通信网络具有结构异常庞大以及干扰情况异常复杂多变的特征,5G的空口资源调度面临着巨大挑战.基站准确获取用户位置不仅是许多应用场景的基本功能需求,还能指导基站在资源调度过程中有意识地规避或者抑制同频干扰,从而进一步提高系统容量.针对传统定位算法存在效率低、计算复杂度高等问题,提出利用介质访问控制(Media Ac...

5G multi-user positioning based on resource scheduling and deep learning

The characteristics of 5G communication network with massive scale and extremely complicated interference can bring great challenges to 5G radio resource scheduling.Getting the location of each user precisely for the base stations is not only a basic functional requirement for many application scenarios,but also can help base station to avoid or suppress co-channel interference consciously in the resource scheduling process to further improve system capacity.In order to cope with the low efficiency and high computational complexity problems in the traditional methods,a novel multi-user position approach is proposed.The proposed approach regards the resource scheduling information naturally generated by the Media Access Control(MAC) layer as fingerprint,and builds a fingerprint data set through decorrelation technique.Moreover,deep learning is used to fit the highly non-linear relationship between MAC layer scheduling information and user location,so as to achieve multi-user positioning without additional signaling overhead.Experimental results show that the proposed method can achieve an accuracy of more than 90% and positioning errors of less than 10 m,which can meet the positioning requirements of 5G.