A Novel Ultra-Massive MIMO BDCM for 6G Wireless Communication Systems

Abstract

In this paper, a novel beam domain channel model (BDCM) for sixth generation (6G) ultra-massive multiple-input multiple-output (MIMO) wireless communication systems is proposed by transforming from an existing geometry-based stochastic model (GBSM). Both the GBSM and BDCM consider the special channel characteristics of ultra-massive MIMO, including spherical wavefront and spatial non-stationarity properties. Steering vectors of the spherical wavefront are derived by higherorder Taylor expansion. By sampling the angle and distance rings, steering matrices of the spherical wavefront can be transformed to unitary matrices. This helps to achieve a perfect transformation from the GBSM to the BDCM in the near-field condition. Meanwhile, it opens up a view of beam domain for channel characteristic analysis to reduce the model complexity. Common statistical properties of the GBSM and BDCM in the near-field and simplified far-field conditions are studied, including spatial cross-correlation functions (SCCFs), temporal autocorrelation functions (TACFs), frequency correlation functions (FCFs), etc. Specific statistical properties, such as the root mean square (RMS) angular spread of the GBSM and RMS beam spread of the BDCM, are studied. Channel capacities of the GBSM and BDCM are investigated and compared with measurement data. It turns out that simulated capacities considering nearfield conditions show good agreements with measured capacities. On the other hand, simulated capacities considering far field conditions show a large discrepancy with measured capacities. This indicates that near-field effects need to be included in the channel modeling and performance evaluation of 6G ultramassive MIMO communication systems.

6G: Technology Evolution in Future Wireless Networks

“…Three low-complexity estimators as Partitioned Sub-array Crosscorrelation Combining (PSCC), Partitioned Sub-array Autocorrelation Combining (PSAC), and power iteration max correlation successive convex approximation are devised in [187] to overcome the challenges of complexity. In [188]…”