Research On Fitting Method And Frequency Allocation Of Cellular Networks Based On Stochastic Geometry

The topology of base station is an important factor affecting network performance.With the continuous development of communication technology,the number and types of base stations are increasing,and the location distribution is becoming more and more complex.The traditional hexagonal mesh model can no longer describe the distribution accurately.Many research conclusions based on traditional cellular network models may not be applicable to the current situation,so the introduction of new models has become an urgent problem to be solved.The model based on stochastic geometry,such as Poisson point process(PPP),attracts a lot of attention.This paper selected four regions of China to analyze,including urban and rural areas.The L function is used to evaluate which point process is best for each region.Results show that the urban region can be modeled as Point cluster process(PCP)and the rural can be modeled as Strauss process(SP).This paper also used coverage probability to compare three fitting methods of Poisson cluster process including minimum contrast,Composite likelihood and Palm likelihood.The conclusion is that the minimum contrast method is best.The irregular distribution of base station makes the traditional frequency reuse scheme no longer suitable for the current situation.In this study,voronin's method was used to model the randomly distributed base station into a graph,and then the graph coloring method was used to color the graph with the least color,so that the color of any connected endpoint was different,that is,the base station with interference was used with different frequency.Then the coverage probability is used to compare the improvement of communication performance before and after frequency reuse.The simulation results show that the performance of frequency multiplexing using graph coloring method is better than that without frequency multiplexing and random frequency multiplexing.