针对分布式massive MIMO系统的导频污染问题，本文提出了一种基于子空间结构化协方差的信道估计算法。该算法利用分布式massive MIMO系统的空间位置特征，基于时延-多普勒-角域信道支持非重叠特性进行导频污染消除。然后，针对所提出算法本文还设计了一套导频分配方案，以降低重叠信道支持对导频污染消除性能的影响。通过理论证明和仿真分析表明，所提算法具有比有代表性的基线方案更好的性能。

A channel estimation algorithm based on subspace structured covariance is proposed in this paper for the pilot contamination problem of distributed massive MIMO systems. The algorithm utilizes the spatial location characteristics of the distributed massive MIMO system and performs the pilot contamination elimination based on the non-overlap characteristics of the delay-Doppler-angle domain channel supports. Then, a pilot assignment scheme is also designed in this paper for the proposed algorithm to reduce the impact of overlapping channel supports on the pilot contamination elimination performance. The theoretical demonstration and simulation analysis show that the proposed algorithm has better performance than the representative baseline scheme.