MU-SHOT-Fi: Self-Supervised Multi-User Wi-Fi Sensing with Source-free Unsupervised Domain Adaptation

2026-05-02

AI Summary

This work addresses the challenges of CSI signal entanglement and unlabeled domain shift in multi-user Wi-Fi sensing caused by environmental dynamics. The authors propose MU-SHOT-Fi, a framework that decouples multiple users in the source domain via permutation-invariant set prediction coupled with Hungarian matching. In the unlabeled target domain, it employs frozen classifier adaptation augmented with spatial self-supervision—implemented through binary rotation prediction—and an occupancy-weighted information maximization strategy to prevent model collapse and enhance cross-domain generalization. A single-user variant, SU-SHOT-Fi, is also introduced, integrating contrastive predictive coding. Experiments demonstrate that the proposed approach substantially recovers activity recognition performance under complex domain shifts—including changes in environment, frequency, and orientation—on the WiMANS and Widar 3.0 datasets, while maintaining accurate user occupancy estimation and mitigating class bias.

Abstract

Deep learning has been widely adopted for WiFi CSI-based human activity recognition (HAR) due to its ability to learn spatio-temporal features in a privacy-preserving and cost-effective manner. However, DL-based models generalize poorly across environments, a challenge amplified in multi-user settings where overlapping activities cause CSI entanglement and domain shifts. Practical deployments often limit access to labeled source data due to privacy constraints, motivating source-free adaptation using only unlabeled target-domain CSI and a pre-trained source model. In this paper, we propose MU-SHOT-Fi, a source-free unsupervised domain adaptation framework for single- and multi-user Wi-Fi sensing. MU-SHOT-Fi employs permutation-invariant set prediction with Hungarian matching during source training, followed by frozen-classifier backbone adaptation in the target domain. To enable stable adaptation without labels, we introduce occupancy-weighted information maximization that prevents model collapse by focusing diversity regularization on likely-occupied slots while excluding the dominant class from marginal entropy. Additionally, we employ binary rotation prediction as spatial self-supervision that exploits CSI frequency-time structure to learn domain-invariant features. For single-user scenarios, we introduce SU-SHOT-Fi by replacing occupancy weighting with standard information maximization and incorporating contrastive predictive coding to exploit temporal consistency. Extensive experiments on the WiMANS and Widar 3.0 datasets across cross-environment, cross-frequency, cross-orientation, and combined domain shifts demonstrate that MU-SHOT-Fi effectively recovers multi-user exact-activity classification performance under large domain shifts while maintaining accurate occupancy estimation and preventing collapse toward dominant classes.

Problem

• Wi-Fi sensing
• human activity recognition
• unsupervised domain adaptation
• multi-user
• source-free

Innovation

• source-free unsupervised domain adaptation
• multi-user Wi-Fi sensing
• permutation-invariant set prediction
• occupancy-weighted information maximization
• spatial self-supervision