Massive MIMO-OFDM Statistical CSI Acquisition with Physical Channel Charting

The deployment of high-dimensional antenna arrays and the growing density of user terminals (UTs) in 6G communication systems have substantially increased the complexity of statistical channel state information (SCSI) acquisition. Meanwhile, advances in positioning technologies have made location-indexed channel fingerprints (CFs) a promising tool for system optimization. In this paper, we introduce the novel concept of channel charting with physical channel fingerprints (PCFs) and propose a low-complexity framework for sCSI acquisition. Firstly, the PCF is defined using a cluster-based geometric stochastic channel model (GBSM), enabling a compact yet expressive representation of channel characteristics using a set of parameters. By collecting PCFs associated with different locations across a cell, a location-indexed channel charting with PCFs can then be constructed. Based on this charting, we develop an efficient algorithm that leverages the PCFs contained in the charting to generate beam domain sCSI, thereby enabling probing-free sCSI acquisition. Simulation results demonstrate that the channel charting-derived sCSI achieves performance comparable to conventional online probing methods while significantly reducing overhead. The acquired sCSI effectively supports essential downstream tasks such as channel estimation.