Coalition Formation for Heterogeneous Federated Learning Enabled Channel Estimation in RIS-assisted Cell-free MIMO

TL;DR

This work addresses the challenge of downlink cascaded channel estimation in RIS-assisted cell-free MIMO by proposing a coalition-formation-guided heterogeneous federated learning framework. It combines DL-based local channel estimation with FL, and leverages DRL (DQN and Qmix) to form coalitions that optimize STN-based model aggregation, while introducing transfer-learning schemes that exploit distance and received power similarities to accelerate convergence. Theoretical guarantees are provided by proving the coalition game is an exact potential game with at least one stable Nash equilibrium, and numerically the approach yields up to ~20% NMSE improvement and about 16% reduction in end-user computational overhead, alongside enhanced data privacy. The proposed framework thus enables scalable, privacy-preserving, and computation-efficient channel estimation in RIS-enabled cell-free networks, with practical impact for 6G-like deployments that require distributed learning and adaptive user grouping.

Abstract

Downlink channel estimation remains a significant bottleneck in reconfigurable intelligent surface-assisted cell-free multiple-input multiple-output communication systems. Conventional approaches primarily rely on centralized deep learning methods to estimate the high-dimensional and complex cascaded channels. These methods require data aggregation from all users for centralized model training, leading to excessive communication overhead and significant data privacy concerns. Additionally, the large size of local learning models imposes heavy computational demands on end users, necessitating strong computational capabilities that most commercial devices lack. To address the aforementioned challenges, a coalition-formation-guided heterogeneous federated learning (FL) framework is proposed. This framework leverages coalition formation to guide the formation of heterogeneous FL user groups for efficient channel estimation. Specifically, by utilizing a distributed deep reinforcement learning (DRL) approach, each FL user intelligently and independently decides whether to join or leave a coalition, aiming at improving channel estimation accuracy, while reducing local model size and computational costs for end users. Moreover, to accelerate the DRL-FL convergence process and reduce computational burdens on end users, a transfer learning method is introduced. This method incorporates both received reference signal power and distance similarity metrics, by considering that nodes with similar distances to the base station and comparable received signal power have a strong likelihood of experiencing similar channel fading. Massive experiments performed that reveal that, compared with the benchmarks, the proposed framework significantly reduces the computational overhead of end users by 16%, improves data privacy, and improves channel estimation accuracy by 20%.

Figures (16)

• Figure 1: RIS-assisted cell-free MIMO network.
• Figure 2: The concept of heterogeneous federated learning networks.
• Figure 3: Aggregation process of heterogeneous federated learning model.
• Figure 4: Composition of models.
• Figure 5: Coalition formation diagram for channel estimation in RIS-assisted cell-free MIMO networks.

Theorems & Definitions (3)

• Remark 1
• Definition 1
• Theorem 1