A Density-Based Clustering Approach for Pilot Assignment in Scalable Cell-Free Massive MIMO

Cell-free massive multiple-input multiple-output (MIMO) is a promising wireless architecture for next-generation networks, but its performance is still strongly affected by pilot contamination, especially in dense and mobile user environments. In this work, we introduce a two-stage pilot assignment framework that leverages structured clustering, systematic refinement, and a channel estimation model that accounts for hardware impairments. In the first stage, we generate a warm-start pilot assignment using an interference-aware hybrid density-based hierarchical clustering method, which groups users based on both spatial proximity and mutual interference. In the second stage, this initial solution is refined using a Tabu Search algorithm whose search space is pruned by a lightweight learning-based mechanism, allowing faster and more effective optimization. In addition, we extend the study to user mobility and evaluate the spectral efficiency of the proposed design under dynamic network conditions. Simulation results confirm that the proposed pilot assignment framework consistently outperforms state-of-the-art pilot assignment schemes under both static and mobile scenarios.