Joint Sparsity Exploitation and Angular Separation for Pilot Contamination Reduction in Massive Multiple-Input Multiple-Output (MIMO)

Massive Multiple-Input Multiple-Output systems suffer from pilot contamination in uplink time-division duplex mode due to the limited availability of orthogonal pilots for channel estimation.This issue becomes more severe in dense networks where pilot reuse across cells is inevitable, leading to degraded channel estimation accuracy and system performance.Efficient pilot allocation strategies can mitigate this problem by minimizing pilot reuse across interfering channels.In this work, users are classified as edge or center users based on their large-scale fading coefficients.The center users are further grouped according to their channel correlation.Orthogonal pilots are assigned to all edge users, while one orthogonal pilot per group is allocated to the center users and reused within the group to reduce the reuse factor and limit contamination.A compressed sensing (CS)-based algorithm is then employed to estimate both the group-wise correlated center-user channels and the edge-user channels, exploiting joint sparsity and per-link sparsity.The CS scheme utilizes an extended discrete Fourier transform basis as the measurement matrix to convert spatialdomain channels into the angular domain, enabling non-overlapping angular supports between groups.Comparative evaluations against state-of-the-art methods demonstrate that the proposed scheme achieves lower channel estimation error and higher throughput, offering a 22.85% improvement in the Cumulative Distribution Function of uplink achievable rate over the SPRS+WGC-PD method at 15 dB Signal-to-Interference-plus-Noise Ratio.