Radio Environment Map Construction Based on Gaussian Process With Positional Uncertainty

Radio Environment Map (REM) is an effective tool for cognitive radio spectrum management. In the process of REM construction, the accurate location of equipment with measurement capability is often difficult to obtain, and inaccurate location information will have a more serious impact on mapping. In this letter, based on variational inference, we propose a Gaussian Process with positional uncertainty (GP-PU) algorithm for constructing a REM within the range of positional uncertainty, and by using the real dataset provided by the 2019 "Huawei" Mathematical Modeling Competition, the effectiveness of the algorithm is verified. In addition, we visualize REM to demonstrate that the GP-PU algorithm can obtain favorable REM. Finally, the simulation results show that the GP-PU algorithm outperforms the traditional method in terms of continuous hierarchical probability score (CRPS) and root mean square error (RMSE).