Mesh-Clustering-Based Radio Maps Construction for Autonomous Distributed Networks

We have proposed a method of the radio map construction using clustering algorithm in our conventional work. The method enables us to accurately predict the radio environment while reducing the registered data size. However, this clustering algorithm has been only applied to the wireless system with fixed transmitter location. Thus, this paper considers the radio maps construction based on the clustering for the autonomous distributed networks that both transmitter and receiver dynamically move. The proposed method classifies the similar average received signal power samples using k-means++. The emulation results clarify that the proposed method can estimate the radio environment with high accuracy while reducing the registered data size compared to the conventional radio map.