m monad-knowledge
community · 5 papers

WiFi CSI Multi-user Activity Sensing

WiFi-based human sensing using channel state information (CSI) is the dominant problem, addressed through deep learning methods including transformer-based architectures. Recent work emphasizes reproducibility, standardization, and benchmarking through unified protocols, libraries, and datasets to enable fair comparison across sensing tasks such as activity recognition and multi-user scenarios.

Members

Papers in this community

  • SenseFi: A library and benchmark on deep-learning-empowered WiFi human sensing 2023 DOI ↗
  • Attention Is All You Need 2017 DOI ↗
  • WiMANS: A Benchmark Dataset for WiFi-based Multi-user Activity Sensing 2025 DOI ↗
  • SDP: A Unified Protocol and Benchmarking Framework for Reproducible Wireless Sensing 2026 DOI ↗
  • CSI-Based NTC Using Ambient WiFi: Channel Selection, Topology Control and Traffic Interference 2026 DOI ↗
Active entities

Methods, problems, datasets, and hardware

Long Short-Term Memory×5 CSI Fingerprinting×4 Convolutional Neural Network×4 Neural Network×3 Transformer / Attention Mechanism×3 Channel State Information×3 Activity Recognition×3 WiFi CSI Sensing×3
Adjacent

Neighbouring communities

WiFi CSI Human Sensing×22 WiFi CSI Human Sensing×2 Knowledge Graph Extraction with LLMs×2