Integrating Intelligent Reflecting Surface into Base Station: Architecture, Channel Model, and Passive Reflection Design

Intelligent reflecting surface (IRS) has emerged as a cost-efficient technique to improve the wireless network’s capacity and performance. Existing works on IRS have mainly considered being deployed in environment dynamically control channels between base station (BS) its served users favor of their communications. In contrast, we propose this paper new integrated IRS-BS architecture by deploying IRSs inside BS’s antenna radome directly reconfigure signal radiation to/from antennas. other words, can be auxiliary passive arrays with real-time reconfigurability equipped at BS enhance communication performance cost-effectively. Since distance array is practically small (in order several tens wavelengths), path loss among them significantly reduced compared conventional much farther away from BS, while IRS’s reflection becomes easier implement. However, resultant near-field channel model also drastically different far-field counterpart for far-away literature. Thus, an element-wise characterize vector each single-antenna user which includes direct (without any reflection) well single double IRS-reflection components. Based model, formulate problem optimize coefficients all elements maximizing uplink sum-rate users. By considering two typical cases with/without perfect state information (CSI) formulated solved efficiently adopting successive refinement method iterative random phase algorithm (IRPA), respectively. Numerical results validate substantial gain over multi-antenna without IRS. Moreover, proposed algorithms outperform benchmark schemes terms sum-rate, IRPA CSI approach upper bound training overhead increases.