Radio Stripe Selection for Improved UE Position Estimation in CF-mMIMO Systems

초록

In this correspondence, we address the problem of radio stripe (RS) selection in cell-free massive multiple-input multiple-output (CF-mMIMO) system based on the instantaneous Cramér-Rao lower bound (CRLB) when fewer RSs than the total available RSs have to be used for an efficient user equipment (UE) position estimation. By leveraging the instantaneous CRLB, the approach proposed in this correspondence minimizes UE position estimation errors and enhances hardware utilization, achieving near-optimal performance with lower computational overhead. The system consists of an unsynchronized single-antenna UE communicating with a network of time- and phase- synchronized RSs distributed across a wide area. To this end, we first formulate UE position estimation problem as a joint estimation of position, clock offset, and phase offset and derive the corresponding maximum-likelihood estimator. We then conduct a Fisher information analysis to gain fundamental insights into the achievable performance and inspect the CRLB numerically. Finally, using these steps, we present a two-stage RSs selection method whose goal is to satisfy the instantaneous CRLB condition, thereby, improving the UE position estimation accuracy. Numerical results demonstrate that the proposed method requires only a small subset of RSs to achieve significant performance gains, while maintaining low computational complexity and delivering superior PEB, spectral efficiency, and runtime performance compared to existing methods.

키워드