Three-Dimensional Channel Power Spectrum Extraction in Massive MIMO for Industrial IoT

In this article, we propose to extract the angle-delay-Doppler power spectrum (ADD-PS) from the acquired uplink (UL) channel states in massive multiple-input-multiple-output systems. Meanwhile, the corresponding power leakages due to finite angle and Doppler resolutions are analyzed in the case of noisy channel state information (CSI) observations at the base station (BS). It is demonstrated that the number of antennas M and the number of observed reference symbols T are complementary in the sense that the amount of power leakages in the ADD-PS will diminish as long as the product of M and T increases. Potential applications of the extracted 3-D channel power spectrum in the industrial Internet of Things (IoT) are also discussed. In particular, the extracted 3-D ADD-PS can be utilized as a powerful fingerprint to facilitate localization in industrial IoT. Thanks to the additional Doppler dimension, it is further demonstrated that the extracted 3-D ADD-PS can also enable the recognition of user moving behavior. We also verify all our findings by carrying out extensive numerical experiments.