Trustworthy Collaboration for WiFi Sensing: Challenges, Methods, and Future Perspectives

WiFi sensing technology leverages environmental and activity information embedded in wireless channel states to achieve understanding of human activities and surroundings, and has been researched and applied in domains such as smart homes, medical rehabilitation, and security surveillance. However, the sensing paradigm based on single-link, single-device, and single-scenario configurations can no longer meet the diverse requirements of real-world deployments, driving research toward collaborative sensing and trustworthy deployment. This paper systematically reviews research progress in the WiFi sensing field from two dimensions: collaboration and trustworthiness. In the collaboration dimension, 4 paradigms are summarized —spatial collaboration, computational collaboration, learning collaboration, and modality collaboration—with analysis of the driving motivations, core methodologies, and capability boundaries of each paradigm. In the trustworthiness dimension, research advances are reviewed across 3 levels: data trustworthiness, model trustworthiness, and security trustworthiness, covering aspects such as data quality control, model robustness enhancement, security attack-defense mechanisms, and fault tolerance design. Based on the above analysis, a unified framework for trustworthy collaborative WiFi sensing is proposed, and future research directions are outlined from the perspectives of integrating physical mechanisms, trustworthy model deployment, privacy-security and collaborative optimization mechanisms, and long-term reliability, providing references for the practical deployment and trustworthy development of WiFi sensing technology.