Generalizable Learning for Massive MIMO CSI Feedback in Unseen Environments

TL;DR

The paper tackles the generalization problem of deep learning–based CSI feedback for FDD massive MIMO by introducing a physics-informed distribution alignment framework. It models cluster-based channel distribution shifts and decomposes them into multi-cluster structure and single-cluster responses, enabling targeted corrections. A novel SVD-based multi-cluster decoupling under the Eckart–Young–Mirsky theorem, combined with a robust cluster-number estimator, enables real-time feedback and resilience to channel estimation errors. The environment-generalizable EG-CsiNet, trained on aligned clusters, achieves more than a 3 dB reduction in generalization error over state-of-the-art methods and demonstrates strong sim-to-real transfer on real-world measurements, indicating significant practical deployment potential.

Abstract

Deep learning is promising to enhance the accuracy and reduce the overhead of channel state information (CSI) feedback, which can boost the capacity of frequency division duplex (FDD) massive multiple-input multiple-output (MIMO) systems. Nevertheless, the generalizability of current deep learning-based CSI feedback algorithms cannot be guaranteed in unseen environments, which induces a high deployment cost. In this paper, the generalizability of deep learning-based CSI feedback is promoted with physics interpretation. Firstly, the distribution shift of the cluster-based channel is modeled, which comprises the multi-cluster structure and single-cluster response. Secondly, the physics-based distribution alignment is proposed to effectively address the distribution shift of the cluster-based channel, which comprises multi-cluster decoupling and fine-grained alignment. Thirdly, the efficiency and robustness of physics-based distribution alignment are enhanced. Explicitly, an efficient multi-cluster decoupling algorithm is proposed based on the Eckart-Young-Mirsky (EYM) theorem to support real-time CSI feedback. Meanwhile, a hybrid criterion to estimate the number of decoupled clusters is designed, which enhances the robustness against channel estimation error. Fourthly, environment-generalizable neural network for CSI feedback (EG-CsiNet) is proposed as a novel learning framework with physics-based distribution alignment. Based on extensive simulations and sim-to-real experiments in various conditions, the proposed EG-CsiNet can robustly reduce the generalization error by more than 3 dB compared to the state-of-the-arts.

Figures (15)

• Figure 1: Proposed EG-CsiNet with strong environment generalizability (left). The physics-based distribution alignment module in EG-CsiNet can effectively address the distribution shift of the cluster-based channel between training and test environments (right).
• Figure 2: Power leakage effect of a two-path cluster in the horizontal angular domain. The AoDs of the two paths are set as $\sin(\phi_{l,1})\sin(\theta_{l,1})=-0.48$ and $\sin(\phi_{l,2})\sin(\theta_{l,2})=-0.35$.
• Figure 3: Numerical validation of cluster orthogonality.
• Figure 4: Detailed structure of proposed EG-CsiNet, where three cluster components are decoupled as an example. The amplitudes of the angular-delay representation for input, output, and intermediate cluster components are also illustrated.
• Figure 5: Illustration of the WAIR-D dataset. The distribution of the number of paths is shown on the left. Example of path parameters is shown on the right, where $\phi^{(\rm az)}/\phi^{(\rm el)}$ denotes the azimuth/elevation AoD.

Theorems & Definitions (4)

• Proposition 1
• Proposition 2
• Theorem 1
• Remark 1