Channel Gain Map Construction based on Subregional Learning and Prediction

--The construction of channel gain map (CGM) is essential for realizing environment-aware wireless communications expected in 6G, for which a fundamental problem is how to predict the channel gains at unknown locations effectively by a finite number of measurements. As using a single prediction model is not effective in complex propagation environments, we propose a subregional learning-based CGM construction scheme, with which the entire map is divided into subregions via data-driven clustering, then individual models are constructed and trained for every subregion. In this way, specific propagation feature in each subregion can be better extracted with finite training data. Moreover, we propose to further improve prediction accuracy by uneven subregion sampling, as well as training data reuse around the subregion boundaries. Simulation results validate the effectiveness of the proposed scheme in CGM construction. To support the largely increased data demands and connection requirements, communication network is becoming more complex in the forthcoming 6G era [1]. This brings challenges to low-complexity network deployment optimization and transmission design [2]. Environment-aware communication provides a promising solution for this challenge, which requires communication-related environment information, also known as the channel knowledge map (CKM) [3], [4], as side information to be exploited when designing the system. In general, the CKM can be presented in terms of a site-specific database, which provides information of concerned channel parameters at given geometric locations. Depending on the particular channel information it conveys, different types of CKM have been studied, including the channel shadowing map [5], channel gain map (CGM) [6], and beam index map [7], etc. Different CKMs can be exploited for different tasks.