Abstract：Inter-cell interference severely limits the network throughput and cell-edge user experience in ultra-dense networks. By using the predict average channel gains from the predicted trajectories and a radio map, with which future resources can be exploited to coordinate interference, balance the traffic load to provide evident performance gain over non-predictive counterparts. However, the existing resource planning studies mainly consider the ideal assumption of predictive information. To investigate the actual performance gain of the resource plan, we develop a trajectory prediction method and a resource planning scheme in ultra-dense networks. The simulation results show that the proposed trajectory prediction method can meet the requirement of resource planning. Furthermore, the proposed resource planning can work well under arbitrary prediction horizons. When the prediction horizon is 3 minutes, compared with the non-predictive interference management method, our resource planning can improve the user satisfaction rate by more than 45%.
Osseiran A, Boccardi F, Braun V, et al.Scenarios for 5G mobile and wireless communications: the vision of the METIS project[J].IEEE Commun. Mag, 2014, 52(5):26-35
[2]
Abou-Zeid H, Hassanein H S.Predictive green wireless access: exploiting mobility and application information[J].IEEE Wireless Commun, 2014, 20(5):92-99
[3]
Yao C, Yang C, Xiong Z.Energy-saving predictive resource planning and allocation[J].IEEE Trans. on Commun., 2016, 64(12):5078-5095
[4]
Yao C, Guo J, Yang C.Achieving high throughput with predictive resource allocation[C]// IEEE GlobalSIP 2016.
[5]
Guo K, Liu T, Yang C, Xiong Z.Interference coordination and resource allocation planning with predicted average channel gains for HetNets[J]. IEEE Access, 2018, 6(6):60137-60151
[6]
Atawia R, Hassanein H S, Abou-Zeid H, et al.Robust content delivery and uncertainty tracking in predictive wireless networks[J].IEEE Trans. on Wireless Commun., 2017, 16(4):2327-2339
[7]
Atawia R, Hassanein H S, Ali N A, et al.Utilization of stochastic modelling for green predictive video delivery under network uncertainties [J].IEEE Trans on Green Commun and Netw, 2018, :1-1
[8]
Chen X, Meriaux F, Valentin S.Predicting a user' s next cell with supervised learning based on channel states[J].IEEE SPAWC, 2013, :-
[9]
Altché, Florent, Arnaud D L F.An LSTM network for highway trajectory prediction[C]// IEEE ITSC 2017.
[10]
Zhang W, Liu Y, Liu T and Yang C.Trajectory prediction with recurrent neural networks for predictive resource allocation[C]//IEEE ICSP 2018.
[11]
Bullen, P.S. Handbook of means and their inequalities [M]. 2003.
[12]
Sutskever I, Vinyals O, Le V Q.Sequence to sequence learning with neural networks[J].NIPS, 2014, :-
[13]
Tomita E, Sutani Y, Higashi T, Wakatsuki M.A simple and faster branch-and-bound algorithm for finding a maximum clique with computational experiments[J].IEICE Trans. Inf. Syst., 2016, 96(6):1286-1298
2019, Vol. 35 
