Massive MIMO-Based Underlay Spectrum Access Under Incomplete and/or Imperfect Channel State Information

We investigate use of massive number of antennas at cognitive base station (BS) in reducing interference caused to primary users (PUs) under incomplete and/or imperfect channel state information (CSI) without deteriorating sum spectral efficiency of cognitive users (CUs). We derive analytical expressions for complement of interference outage probability with incomplete and/or imperfect CSI. These novel expressions provide ballpark number for the amount of back-off required to meet interference constraint. We deduce new expressions for sum spectral efficiency of CUs with MR beamforming under incomplete and/or imperfect CSI. We prove that under imperfect CSI, by deploying more antennas at cognitive BS, interference outage probability can be reduced while keeping sum spectral efficiency of CUs fixed. And larger number of PUs can be accommodated in the network while keeping interference outage probability at PUs and sum spectral efficiency of CUs fixed. Furthermore, outage probability reduces as number of channels
${S}$
to which CSI is available increases, since the constraint is violated less often. And by deploying more antennas at cognitive BS, sum spectral efficiency of CUs can be kept fixed even if
${S}$
decreases. Impact of spatial correlation on outage probability at PUs and sum spectral efficiency of CUs is also elucidated.