Conditional Prior-based Non-stationary Channel Estimation Using Accelerated Diffusion Models

2025-09-18

AI Summary

To address channel non-stationarity induced by user mobility in urban microcell (UMi) scenarios, this paper proposes a history-aware conditional prior diffusion model for wireless channel estimation. Methodologically, it integrates geometrically spaced diffusion scheduling, SNR-matched initialization, temporal encoding, and cross-temporal attention, augmented with feature-level modulation and temporal self-conditioned feedback to explicitly capture time-varying channel dynamics. Evaluated on the 3GPP UMi benchmark, the model achieves significantly lower normalized mean square error (NMSE) than conventional LMMSE and GMM estimators, as well as deep learning baselines (LSTM, LDAMP), maintaining high accuracy across the full SNR range and exhibiting superior robustness at high SNR. The core contribution lies in embedding physics-informed temporal priors directly into the diffusion process—balancing modeling fidelity and computational efficiency—and achieving, for the first time, high-fidelity, low-iteration deep channel estimation under non-stationary conditions.

Technology Category

• Intelligent Robots: State Estimation
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Web Mining and Content Analysis: Content-based information diffusion

Abstract

Wireless channels in motion-rich urban microcell (UMi) settings are non-stationary; mobility and scatterer dynamics shift the distribution over time, degrading classical and deep estimators. This work proposes conditional prior diffusion for channel estimation, which learns a history-conditioned score to denoise noisy channel snapshots. A temporal encoder with cross-time attention compresses a short observation window into a context vector, which captures the channel's instantaneous coherence and steers the denoiser via feature-wise modulation. In inference, an SNR-matched initialization selects the diffusion step whose marginal aligns with the measured input SNR, and the process follows a shortened, geometrically spaced schedule, preserving the signal-to-noise trajectory with far fewer iterations. Temporal self-conditioning with the previous channel estimate and a training-only smoothness penalty further stabilizes evolution without biasing the test-time estimator. Evaluations on a 3GPP benchmark show lower NMSE across all SNRs than LMMSE, GMM, LSTM, and LDAMP baselines, demonstrating stable performance and strong high SNR fidelity.

Problem

• Estimating non-stationary wireless channels in urban microcells
• Denoising noisy channel snapshots with history-conditioned diffusion models
• Improving channel estimation accuracy across varying SNR conditions

Innovation

• Conditional prior diffusion for channel estimation
• Temporal encoder with cross-time attention
• SNR-matched initialization with shortened schedule