Automatic Clustering for Improved Radio Environment Maps in Distributed Applications

Wireless communication greatly contributes to the evolution of new technologies, such as Internet Things (IoT) and edge computing. The generation networks, including 5G 6G, provide several connectivity advantages for multiple applications, smart health systems cities. Adopting wireless technologies in these applications is still challenging due factors mobility heterogeneity. Predicting accurate radio environment maps (REMs) essential facilitate improve resource utilization. construction REMs through prediction reference signal received power (RSRP) can be useful densely distributed However, predicting an RSRP complex intervention aspects. Given fact that propagation environments different a specific area interest, estimation common path loss exponent entire produces errors constructed REM. Hence, it necessary use automatic clustering distinguish between by grouping locations exhibit similar characteristics. This leads better characteristics other within same cluster. Therefore, this work, we propose using Kriging technique, conjunction with approach, order accuracy prediction. In fact, adopt K-means (KMC) enhance estimation. We dataset test proposed model set comparative studies. results showed approach provides significant capabilities constructing REM, gain about 3.3 dB terms root mean square error compared case without clustering.