利用终端位置时空转移概率预测通讯基站服务用户规模

发布时间：2019-07-29 17:49

【摘要】：基站服务用户数的预测对通信基站的空间位置选取、通讯服务带宽的配置优化等有重要作用,并为城市公共安全管理方面中的人群聚集预警与群体事件预防提供辅助决策支持。本文利用海量手机轨迹数据,运用时空转移概率定量化描述城市内不同区域间人群流动的时空特性,结合马尔可夫链和贝叶斯定理,构建手机用户群体在基站间的时空转移概率模型,并以此提出一种城市区域尺度上的基站服务用户数预测方法。利用湖北省某城市长达30 d的手机轨迹数据进行模型训练与预测方法验证,实验结果表明:在时间粒度为60 min时,本文提出的方法对8-22时各时段基站服务用户数预测准确率都大于94.8%;在不同时间粒度下对比本文方法、Castro模型、移动平均法,发现在时间粒度大于20 min时,本文方法预测准确率高于另外两种方法。
[Abstract]:The prediction of the number of base station service users plays an important role in the spatial location selection of communication base stations and the optimization of communication service bandwidth configuration, and provides auxiliary decision support for crowd aggregation early warning and group event prevention in urban public safety management. In this paper, based on the massive mobile phone trajectory data, the temporal and spatial transfer probability is used to quantitatively describe the temporal and spatial characteristics of crowd flow between different regions of the city, combined with Markov chain and Bayesian theorem, the temporal and spatial transfer probability model of mobile phone user groups between base stations is constructed, and a prediction method of the number of base station service users on the urban regional scale is proposed. The model training and prediction method is verified by using the 30-day mobile phone trajectory data of a city in Hubei Province. the experimental results show that when the time granularity is 60 min, the accuracy of the method proposed in this paper is more than 94.8% for the number of base station service users in each period of 8-22:00. Compared with the proposed method, Castro model and moving average method under different time granularity, it is found that when the time granularity is more than 20 min, the prediction accuracy of this method is higher than that of the other two methods.
【作者单位】：武汉大学测绘遥感信息工程国家重点实验室;地球空间信息技术协同创新中心;中国移动通信集团湖北有限公司业务支撑中心;
【基金】：国家自然科学基金项目(41231171、41371420) 湖北省青年英才开发计划项目武汉大学自主科研项目拔尖创新人才类资助项目(2042015KF0167)
【分类号】：O213

【相似文献】