Leveraging Beam-Level RSRP of 5G mmWave Signal for Indoor Passive Crowd Counting

Crowd counting estimation are crucial for applications such as intrusion detection, event crowd control, and retail staffing optimization. Traditional visual-based systems, while effective, are limited by challenges in low-light conditions, privacy concerns, and restricted field-of-view. RF-based approaches offer a non-intrusive alternative but struggle with defining coverage zones and update rates. This work explores the use of 5G mmWave signal to address these issues, leveraging beam-level reference signal received power (RSRP) for crowd monitoring. Using a fingerprinting approach and K-nearest neighbor classifier with additional beam-level RSRP process, this work comprehensively evaluates the crowd counting performance with the crowd up to five walking individuals in indoor office. The results demonstrated the feasibility of using commercial 5G mobile communication system for passive indoor crowd counting.