Generative MIMO Beam Map Construction for Location Recovery and Beam Tracking

2025-11-21

AI Summary

This paper addresses localization and beam tracking from sparse CSI sequences without explicit position labels. We propose a generative self-supervised framework that employs a dual-scale recurrent-convolutional encoder to jointly extract spatiotemporal and angular-domain features, constructing a learnable low-dimensional RF-map latent space. A diffusion-based generative decoder is then used to reconstruct MIMO beam patterns with high fidelity, overcoming the limitations of conventional Gaussian priors. Our key innovations include implicit encoding of positional information and joint modeling of spatiotemporal dependencies across neighboring samples and angular correlations. Experiments demonstrate that, under non-line-of-sight conditions, our method improves localization accuracy by over 30% compared to Kalman filtering, while achieving a 20% gain in channel capacity. Crucially, it significantly reduces reliance on large-scale, precisely annotated datasets.

Technology Category

• Intelligent Robots: Localization, Mapping, and Navigation
• Machine Learning: Deep Generative Models & Autoencoders
• Computer Vision: Motion & Tracking

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Location- and context-aware Web and WoT applications and services
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search

Abstract

Machine learning (ML) has greatly advanced data-driven channel modeling and resource optimization in wireless communication systems. However, most existing ML-based methods rely on large, accurately labeled datasets with location information, which are often difficult and costly to obtain. This paper proposes a generative framework to recover location labels directly from sequences of sparse channel state information (CSI) measurements, without explicit location labels for radio map construction. Instead of directly storing raw CSI, we learn a compact low-dimensional radio map embedding and leverage a generative model to reconstruct the high-dimensional CSI. Specifically, to address the uncertainty of sparse CSI, a dual-scale feature extraction scheme is designed to enhance feature representation by jointly exploiting correlations from angular space and across neighboring samples. We develop a hybrid recurrent-convolutional encoder to learn mobility patterns, which combines a truncation strategy and multi-scale convolutions in the recurrent neural network (RNN) to ensure feature robustness against short-term fluctuations. Unlike conventional Gaussian priors in latent space, we embed a learnable radio map to capture the location information by encoding high-level positional features from CSI measurements. Finally, a diffusion-based generative decoder reconstructs the full CSI with high fidelity by conditioning on the positional features in the radio map. Numerical experiments demonstrate that the proposed model can improve localization accuracy by over 30% and achieve a 20% capacity gain in non-line-of-sight (NLOS) scenarios compared with model-based Kalman filter approaches.

Problem

• Recover location labels from sparse CSI measurements without explicit labels
• Construct radio maps without storing raw channel state information
• Enhance localization accuracy and capacity in non-line-of-sight scenarios

Innovation

• Generative framework recovers location from sparse CSI
• Dual-scale feature extraction enhances angular correlation representation
• Diffusion-based decoder reconstructs high-fidelity CSI using positional features