A low complexity distributed cluster based algorithm for spatial prediction

Radio Environment Maps (REM) can be an essential tool for numerous applications in future 5G wireless networks. In this paper, we employ a popular geo-statistical method called ordinary kriging to estimate the REM of an area covered by a eNodeB equipped with multiple antennas. Wireless sensors are distributed over the area of interest and adaptive clusters of sensors are arranged in order to improve the quality of the estimation. In this paper, we modify the distributed clustering algorithm proposed in a previous work to reduce the complexity of kriging prediction. Simulations are performed to detail the cluster formation technique and to analyze the performance in comparison with centralized and classical interpolation methods. The computational complexity is verified in terms of the number of message exchanges among the sensor nodes. Simulation results demonstrate that clusters are formed by an average of 5 sensor nodes.

[1]  Jordi Pérez-Romero,et al.  On the use of radio environment maps for interference management in heterogeneous networks , 2015, IEEE Communications Magazine.

[2]  N. Cressie.  Fitting variogram models by weighted least squares , 1985 .

[3]  Sana Ben Jemaa,et al.  Harvesting MDT data: Radio environment maps for coverage analysis in cellular networks , 2013 .

[4]  Baltasar Beferull-Lozano,et al.  Field estimation in wireless sensor networks using distributed kriging , 2012, 2012 IEEE International Conference on Communications (ICC).

[5]  Georgios B. Giannakis,et al.  Channel Gain Map Tracking via Distributed Kriging , 2011, IEEE Transactions on Vehicular Technology.

[6]  Hossam S. Hassanein,et al.  Optimal predictive resource allocation: Exploiting mobility patterns and radio maps , 2013, 2013 IEEE Globecom Workshops (GC Wkshps).

[7]  Mehrzad Malmirchegini,et al.  On the Spatial Predictability of Communication Channels , 2012, IEEE Transactions on Wireless Communications.

[8]  Tuna Tugcu,et al.  Radio environment map as enabler for practical cognitive radio networks , 2013, IEEE Communications Magazine.

[9]  Baltasar Beferull-Lozano,et al.  Distributed Clustering Algorithm for Spatial Field Reconstruction in Wireless Sensor Networks , 2015, 2015 IEEE 81st Vehicular Technology Conference (VTC Spring).

[10]  Richardo A. Olea,et al.  Semivariogram modeling by weighted least squares , 1996 .

[11]  Eryk Dutkiewicz,et al.  Special Interest Group Cognitive Radio in 5 G White Paper Novel Spectrum Usage Paradigms for 5 , 2014 .

[12]  Margaret A. Oliver,et al.  A tutorial guide to geostatistics: Computing and modelling variograms and kriging , 2014 .

[13]  X. Emery.  The kriging update equations and their application to the selection of neighboring data , 2009 .