A Survey on Wi-Fi Sensing Generalizability: Taxonomy, Techniques, Datasets, and Future Research Prospects

TL;DR

This survey provides a comprehensive and structured review of over 200 papers published since 2015, categorizing them according to the Wi-Fi sensing pipeline: experimental setup, signal preprocessing, feature learning, and model deployment, and discusses emerging trends and future directions.

Abstract

Wi-Fi sensing has emerged as a powerful non-intrusive technology for recognizing human activities, monitoring vital signs, and enabling context-aware applications using commercial wireless devices. However, the performance of Wi-Fi sensing often degrades when applied to new users, devices, or environments due to significant domain shifts. To address this challenge, researchers have proposed a wide range of generalization techniques aimed at enhancing the robustness and adaptability of Wi-Fi sensing systems. In this survey, we provide a comprehensive and structured review of over 200 papers published since 2015, categorizing them according to the Wi-Fi sensing pipeline: experimental setup, signal preprocessing, feature learning, and model deployment. We analyze key techniques, including signal preprocessing, domain adaptation, meta-learning, metric learning, data augmentation, cross-modal alignment, federated learning, and continual learning. Furthermore, we summarize publicly available datasets across various tasks, such as activity recognition, user identification, indoor localization, and pose estimation, and provide insights into their domain diversity. We also discuss emerging trends and future directions, including large-scale pretraining, integration with multimodal foundation models, and continual deployment. To foster community collaboration, we introduce the Sensing Dataset Platform (SDP) (http://www.sdp8.org/) for sharing datasets and models. This survey aims to serve as a valuable reference and practical guide for researchers and practitioners dedicated to improving the generalizability of Wi-Fi sensing systems.

Figures (17)

• Figure 1: Wi-Fi signals emitted from a transmitter propagate through both static environmental structures and dynamic human bodies before reaching the receiver. Human presence and movements alter the signal propagation paths, leading to measurable changes in the received signal. These variations can be leveraged for Wi-Fi sensing to interpret human movements.
• Figure 2: Wi-Fi sensing generalizability is primarily hindered by three key factors, i.e., device heterogeneity, human body diversity, and environmental diversity.
• Figure 3: Growth of research in Wi-Fi sensing generalizability: from a handful of studies between year of 2015 and 2018 to a surge of publications since 2019. We employed a linear regression to fit the growth trend. In the figure, the $R^2$ score indicates the correlation between the estimated results and the ground truth, where a value closer to 1 denotes a higher degree of model fitting. (We selected one representative paper per year based on the highest citation count, ensuring each choice employs a technical approach distinct from those featured in previous years.)
• Figure 4: We systematically review and categorize nearly a decade of research on Wi-Fi sensing generalization, covering over 200 publications since 2015. Distinct from previous surveys, we organize the literature along a four-stage sensing pipeline—experimental setup, signal preprocessing, feature learning, and model deployment.
• Figure 5: In the experimental setup stage, Wi-Fi generalization can be enhanced by distributing antennas to mitigate the impact of user orientation, deploying devices more widely or optimizing their placement to increase coverage, and collecting more diverse datasets to support the training of more robust and generalizable models.