面向5G需求的人群流量预测模型研究

doi:10.11959/j.issn.1000-436x.2019042

通信学报 ›› 2019, Vol. 40 ›› Issue (2): 1-10.doi: 10.11959/j.issn.1000-436x.2019042

• 专题：5G与AI • 下一篇

面向5G需求的人群流量预测模型研究

胡铮,袁浩,朱新宁,倪万里

北京邮电大学网络与交换技术国家重点实验室，北京 100876

修回日期:2019-02-14 出版日期:2019-02-01 发布日期:2019-03-04
作者简介:胡铮（1980- ），男，贵州贵阳人，博士，北京邮电大学副教授，主要研究方向为行为认知与智能系统。|袁浩（1994- ），男，河北衡水人，北京邮电大学硕士生，主要研究方向为交通大数据、轨迹数据挖掘与分析等。|朱新宁（1970- ），女，上海人，博士，北京邮电大学副教授，主要研究方向为时空大数据挖掘。|倪万里（1995- ），男，重庆人，北京邮电大学博士生，主要研究方向为边缘计算。
基金资助:
国家自然科学基金资助项目(61421061);海南省重大科技计划基金资助项目(ZDKJ201808)

Research on crowd flows prediction model for 5G demand

Zheng HU,Hao YUAN,Xinning ZHU,Wanli NI

State Key Laboratory of Networking and Switching Technology,Beijing University of Posts and Telecommunications,Beijing 100876,China

Revised:2019-02-14 Online:2019-02-01 Published:2019-03-04
Supported by:
The National Natural Science Foundation of China(61421061);The Major Science and Technology Plan Project of Hainan Provinc(ZDKJ201808)

摘要/Abstract

摘要：

5G 网络中超密集基站的部署规划、多维资源管理、活跃/休眠切换等方面都依赖于对区域内用户数量的准确预测。针对这一需求，提出了一种基于移动网络用户位置信息的区域人群流量预测的深度时空网络模型。通过建模不同尺度的时空依赖关系，融合各种外部特征信息，并以短时局部流量信息降低对实时全局信息传输的要求，实现了城市范围的区域人群流量预测，对提高5G网络性能具有重要意义。通过基于呼叫详单数据的区域人群流量预测实验表明，与现有流量预测模型相比，所提模型具有更高的预测精度。

关键词: 5G网络, 人群流量预测, 深度神经网络, 时空数据挖掘

Abstract:

The deployment and planning for ultra-dense base stations,multidimensional resource management,and on-off switching in 5G networks rely on the accurate prediction of crowd flows in the specific areas.A deep spatial-temporal network for regional crowd flows prediction was proposed,by using the spatial-temporal data acquired from mobile networks.A deep learning based method was used to model the spatial-temporal dependencies with different scales.External factors were combined further to predict citywide crowd flows.Only data from local regions was applied to model the closeness of properties of the crowd flows,in order to reduce the requirements for transmitting the globe data in real time.It is of importance for improving the performance of 5G networks.The proposed model was evaluated based on call detail record data set.The experiment results show that the proposed model outperforms the other prediction models in term of the prediction precision.

Key words: 5G networks, crowd flows prediction, deep neural networks, spatial-temporal data mining

中图分类号:

TP393

胡铮,袁浩,朱新宁,倪万里. 面向5G需求的人群流量预测模型研究[J]. 通信学报, 2019, 40(2): 1-10.

Zheng HU,Hao YUAN,Xinning ZHU,Wanli NI. Research on crowd flows prediction model for 5G demand[J]. Journal on Communications, 2019, 40(2): 1-10.

图/表 12

图1

表1

符号标记及其含义"

符号标记	含义
M_i	第i个待预测的区域内的基站集合
(i,j)	城市分区结果中的第i行、第j列的区域
B_(i,j)	网格(i,j)内的基站集合
p	要预测的区域数目
F_t(i,j)	第t个时间段的网格(i,j)内的流量
F_t	第t个时间段的流量图
S_t	第t个时间段的区域真实流量向量
${\hat{S}}_{t}$	第t个时间段的区域预测流量向量
c,pr,tr	临近、周期和趋势时间层的取样周期
l_c,l_pr,l_tr	临近、周期和趋势时间层的取样长度
G	区域距离图
S_ST	时空模块的输出向量

表1

图2

表2

图3

图4

表3

图5

图6

图7

图8

图9

参考文献 19

[1]	CHEN M , QIAN Y , HAO Y ,et al. Data-driven computing and caching in 5G networks:architecture and delay analysis[J]. IEEE Wireless Communications, 2018,25(1): 70-75.
[2]	CHEN N C , XIE W , WELSCH R E ,et al. Comprehensive predictions of tourists' next visit location based on call detail records using machine learning and deep learning methods[C]// IEEE International Congress on Big Data (BigData Congress). 2017: 1-6.
[3]	FENG M , MAO S , JIANG T . Base station on-off switching in 5G wireless networks:approaches and challenges[J]. IEEE Wireless Communications, 2017,24(4): 46-54.
[4]	OUGHTON E , FRIAS Z , RUSSELL T ,et al. Towards 5G:scenario-based assessment of the future supply and demand for mobile telecommunications infrastructure[J]. Technological Forecasting and Social Change, 2018,S0040162517313525.
[5]	KUMAR S V . Traffic flow prediction using kalman filtering technique[J]. Procedia Engineering, 2017,187: 582-587.
[6]	LI X , PAN G , WU Z ,et al. Prediction of urban human mobility using large-scale taxi traces and its applications[J]. Frontiers of Computer Science in China, 2012,6(1): 111-121.
[7]	MATIAS M , LUIS , GAMA ,et al. Predicting taxi-passenger demand using streaming data[J]. IEEE Transactions on Intelligent Transportation Systems, 2013,14(3): 1393-1402.
[8]	TONG Y , CHEN Y , ZHOU Z ,et al. The simpler the beter:a unified approach to predicting original taxi demands based on large-scale online platforms[C]// ACM SIGKDD International Conference on Knowledge Discovery ＆ Data Mining. ACM, 2017: 1653-1662.
[9]	WU F , WANG H , LI Z . Interpreting traffic dynamics using ubiquitous urban data[C]// ACM Sigspatial International Conference on Advances in Geographic Information Systems. 2016,69(3): 1-4.
[10]	YAO H , WU F , KE J ,et al. Deep multi-view spatial-temporal network for taxi demand prediction[C]// AAAI Conference on Artificial Intelligence. 2018: 1-8.
[11]	LIU Y , ZHENG H , FENG X ,et al. Short-term traffic flow prediction with Conv-LSTM[C]// International Conference on Wireless Communications ＆ Signal Processing. 2017: 1-6.
[12]	POLSON N G , SOKOLOV V O . Deep learning for short-term traffic flow prediction[J]. Transportation Research Part C:Emerging Technologies, 2017,79: 1-17.
[13]	ZHANG J , ZHENG Y , QI D . Deep spatio-temporal residual networks for citywide crowd flows prediction[C]// AAAI. 2017: 1655-1661.
[14]	ZHANG J , ZHENG Y , QI D ,et al. DNN-based prediction model for spatio-temporal data[C]// Proceedings of the 24th ACM SIGSPATIAL International Conference on Advances in Geographic Information Systems. 2016:92.
[15]	HUANG C , ZHANG J , ZHENG Y ,et al. DeepCrime:attentive hierarchical recurrent networks for crime prediction[C]// The 27th ACM International Conference. 2018: 1423-1432.
[16]	ZHANG J , ZHENG Y , QI D ,et al. Predicting citywide crowd flows using deep spatio-temporal residual networks[J]. Artificial Intelligence, 2018:S0004370218300973.
[17]	GROVER A , LESKOVEC J . Node2vec:scalable feature learning for networks[C]// ACM SIGKDD Conference on know Ledge Discovery and Data Mining. 2016: 855-864.
[18]	WU W , WANG Y , GOMES J B ,et al. Oscillation resolution for mobile phone cellular tower data to enable mobility modelling[C]// IEEE International Conference on Mobile Data Management.IEEE Computer Society. 2014: 321-328.
[20]	XIAOLEI M , ZHUANG D , ZHENGBING H ,et al. Learning traffic as images:a deep convolutional neural network for large-scale transportation network speed prediction[J]. Sensors, 2017,17(4): 818-833.
[19]	SHAD S A , CHEN E , BAO T . Cell oscillation resolution in mobility profile building[J]. International Journal of Computer Science Issues, 2012,9(3).

类别	特征描述
时间信息	每周的第几天；每天的第几个小时
节假日信息	是否节假日（0,1：节假日记为0，否则为1）
	节假日假期的长度
气象信息	天气状况
	最高和最低气温
	风力
	湿度
空气质量信息	AQI级别

算法模型	工作日	周末
HA模型	112.4	141.3
ARIMA模型	95.6	119.1
VAR模型	79.5	101.5
卡尔曼滤波模型	84.5	99.5
LSTM-12模型	65.3	77.4
LSTM-48模型	63.1	74.1
ST-image模型	68.5	89.5
DeepST模型	50.5	57.4
ST-ResNet模型	46.5	52.1
本文模型变种1	48.8	55.4
本文模型变种2	43.3	49.9
本文模型变种3	47.1	53.9

面向5G需求的人群流量预测模型研究

Research on crowd flows prediction model for 5G demand

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 19

相关文章 8

Metrics

推荐阅读 0

[1]	林云, 徐怀韬, 王森, 张思成, 庄龙. 基于特征融合的通信语音干扰效果客观评估[J]. 通信学报, 2023, 44(3): 105-116.
[2]	熊金波, 周永洁, 毕仁万, 万良, 田有亮. 边缘协同的轻量级隐私保护分类框架[J]. 通信学报, 2022, 43(1): 127-137.
[3]	王哲, 李陶深, 葛丽娜, 张桂芬, 吴敏. 基于深度学习的传感云sink节点最优能效SWIPT波束成形设计[J]. 通信学报, 2021, 42(7): 176-188.
[4]	傅友华, 陈栋. 混合智能反射表面结构辅助的毫米波通信信道估计[J]. 通信学报, 2021, 42(10): 189-196.
[5]	吴伟华,柴冠华,杨清海,刘润滋. 面向不确定CSI随机接入网络的深度稳健资源分配[J]. 通信学报, 2020, 41(7): 29-37.
[6]	张思成,林云,涂涯,ShiwenMao. 基于轻量级深度神经网络的电磁信号调制识别技术[J]. 通信学报, 2020, 41(11): 12-21.
[7]	桂冠,王禹,黄浩. 基于深度学习的物理层无线通信技术：机遇与挑战[J]. 通信学报, 2019, 40(2): 19-23.
[8]	高洪元,杜亚男,张世铂,刘丹丹. 能量采集认知无线电的量子蝙蝠最优合作策略[J]. 通信学报, 2018, 39(9): 10-19.