REM Reconstruction Based on Optimized Regression Scheme

In order to monitor the electromagnetic environment and signal distribution accurately in wireless communication networks, it is essential to reconstruct the radio electromagnetic map (REM). In this paper, an improved gaussian process regression (GPR) scheme is proposed to realize the REM reconstruction based on the optimization algorithm. The simulation results show that the scheme excels in interpolating a large area with an RMSE of 2.3041, even though the sampling nodes account for only 1%. Besides, it has the advantages of fast convergence of errors. The scheme is a good candidate for reconstructing REM under complex conditions and with little sampling nodes, which helps to realize the quantitative description of the signal distribution.