A Comprehensive Review of Wi-Fi-Based Indoor Human Activity Detection

Abstract

Driven by the rising demand for privacy-friendly, unobtrusive indoor monitoring, Wi-Fi passive human sensing has developed into a widely studied technical paradigm. Compared with vision-based and wearable sensing techniques, Wi-Fi sensing exhibits unique advantages including robust illumination, privacy protection, wall penetration, device-free operation, and low hardware cost. Driven by advanced deep learning, this field has advanced substantially over the past decade, evolving from coarse-level activity classification to fine-grained tasks including 3D human skeleton estimation, dense body pose matching, and multi-user activity sensing. However, several fundamental bottlenecks remain unresolved, including severe performance degradation in cross-domain and cross-environment scenarios, insufficient spatial resolution for motion sensing, limited capability for multi-user signal disentanglement, and the lack of standardized benchmark datasets. This survey provides a systematic overview of cutting-edge research on Wi-Fi-enabled indoor human activity detection. We classify mainstream technologies along three dimensions: signal processing pipelines, learning paradigms, and application granularity, and further dissect core challenges including environmental adaptability, data scarcity, and system scalability. Finally, we highlight promising future research directions covering domain generalization, cross-modal foundation model distillation, multi-user collaborative sensing, and real-time on-device deployment.