Indoor personnel detection and tracking of millimeter-wave radar based on improved DBSCAN algorithm

With the progress of technology and the enhancement of social demand for privacy protection, optical monitoring equipment has gradually caused public concern. In contrast, millimeter-wave(mmWave) radar monitoring has been rapidly developed because of its superiority in privacy. However, the indoor e...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Engineering Research Express Jg. 7; H. 2; S. 25220 - 25235
Hauptverfasser: Zhou, Fang, Gao, Yuan, Li, Andong, Xing, Mengdao
Format: Journal Article
Sprache:Englisch
Veröffentlicht: IOP Publishing 30.06.2025
Schlagworte:
clustering algorithm
indoor tracking
mmWave radar
multi-frame aggregation
ISSN:2631-8695, 2631-8695
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:With the progress of technology and the enhancement of social demand for privacy protection, optical monitoring equipment has gradually caused public concern. In contrast, millimeter-wave(mmWave) radar monitoring has been rapidly developed because of its superiority in privacy. However, the indoor environment is relatively complex, and traditional density-based clustering algorithms perform poorly in accurate tracking. The point cloud data generated from indoor scenario echoes collected by mmWave radar is relatively sparse and accompanied by noise points, which significantly affects tracking performance. In this paper, we propose an improved DBSCAN clustering algorithm that uses a multi-frame aggregation method to suppress multipath effects and eliminate ‘false targets’. It is combined with the extended Kalman filter(EKF) algorithm to form a complete system. In our system, the raw data collected by mmWave radar is processed by fast fourier transform(FFT), static clutter removal and constant false alarm rate(CFAR) to obtain point cloud data. Since the density of point cloud data greatly affects the performance of clustering algorithms, we use multi-frame aggregation method to process the point cloud data to increase its density. Accurate indoor personnel tracking is then achieved through clustering and extended Kalman filtering, and the tracking error is within 0.1 m.
Bibliographie:ERX-107030.R2
ISSN:2631-8695
2631-8695
DOI:10.1088/2631-8695/adcc7a