Spatial Correlation in Indoor Massive MIMO: Measurements and Ray Tracing

This letter investigates spatial correlation properties of radio-frequency massive multiple-input-multiple-output (MIMO) channels. The correlation between channel vectors at the receive side is studied for the interreceiver distances up to several wavelengths. This is an important property of 5G wireless networks and its accurate prediction is desirable in many applications. Measured and simulated channels are compared using an alternative formulation of the channel correlation function, the argument of which is the distance between any two receivers in proximity of a selected region. The measurements were conducted using a massive MIMO base station (BS) testbed and virtual arrays of receivers, positioned with a robotic system. The simulations were performed using the ray-tracing (RT) technique in a simplified model of an indoor environment, augmented with stochastic geometry elements. In addition, a line-of-sight (LOS) channel model was derived from the RT results, and the role of the scattered power on the correlation was evaluated. The results show that the RT model predicts the correlation with an error not exceeding 10%. The variation of the correlation function profile with increasing receiver to BS separation distance is also captured by the RT model. The simulated LOS model predictions, which account solely for direct propagation paths, were found to significantly underestimate the correlation at the subwavelength distance, and overestimate it at larger distances.

Automated summary

The letter investigates spatial correlation properties of radio-frequency massive MIMO channels, focusing on the correlation between channel vectors at the receive side. The study shows that the ray-tracing model accurately predicts correlation within several wavelengths of interreceiver distances, capturing the variation of the correlation function profile with increasing receiver to BS separation distance.