Accurate Analysis Of Coverage And Interference In TD-LTE Power Wireless Private Networks

The smart grid has a huge significance for promoting the sustainable development of the world economyand society,and advanced power communications can ensure successful operating of the smart grid,and the TD-LTE power wireless private network is the key and foundation for the power communications in China.Network planning and optimization of TD-LTE power wireless private networks requires accurate analysis of coverage and interference.We have done the following work for these obj ectives.(1)Accurate radio propagation models are critical for coverage analysis.We develop a ray tracing based coverage analysis scheme.This method uses spatial clustering to divide the coverage area in a dense urban environment into several geographical scenarios,and then uses the information on terrain and the on-ground objects,e.g.buildings provided by the 3D digital map,and the scenario-oriented ray tracing propagation model to calculate the geographical distribution of cell signals.However,analysis accuracy suffers from imprecision in buildings' dielectric coefficients in the model.Thus,we model calibration of the direct,reflection and diffraction parameters in different scenarios used by the model as a multiobjective optimization problem,and employ the driving-test data and the multiobjective evolutionary algorithm to calibrate these parameters so as to improve the model accuracy.(2)Accurate estimation of cell coverage based on measurerment data is vital to planning and deploying of TD-LTE power wireless networks in regions where exist coverage holes.We propose a method of estimating cell coverage by using the cell coverage characteristics map.This method collects network measurement data such as driving-test data to build a cell coverage characteristic map that shows the frequency,RSRP(Reference Signal Receiving Power)and PCI(Physical Cell Identifier)in the grid positions over the coverage area,then cleans and compensates the errors of the outliers in the map,and discerns the coverage border line for each cell.The border line ensures integrity of coverage characteristic data,and represent the effective estimation of cell coverage.(3)SINR in drive-testing data precisely records the sum of real intranet and extranet interference received in the measuring spots along the driving-testing paths.In order to take the real SINR values collected in a limited number of measuring positions in network to evaluate SINR distribution over the network,we present an approach of SINR predication by tensor decomposition.This approach divides the coverage area into grids,and employs the drive-testing data to model the SINR interference as a multi-dimensional tensor,which describes the spatial correlation among the SINR values,grid locations,serving and neighbor cells,and terrain features in the coverage area.By means of tensor decomposition,the missing SINR values in the tensor are supplemented,thus the SINR values in the non-measuring positions are predicted.We validate the effectiveness of above-mentioned research by using the network parameters and measurement data collected from the practical TD-LTE power wireless private network deployed in Suzhou,China.