Human body posture recognition method based on FMCW radar signal

The invention provides a human body posture recognition method based on an FMCW radar signal. The method comprises the steps of firstly carrying out the discrete Fourier transform (FFT) of data collected by an FMCW radar, and obtaining a target distance, a target speed and a target angle, secondly,...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: ZHENG HAIFENG, FENG XINXIN, LI WENLONG, NIE JUNJUN
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The invention provides a human body posture recognition method based on an FMCW radar signal. The method comprises the steps of firstly carrying out the discrete Fourier transform (FFT) of data collected by an FMCW radar, and obtaining a target distance, a target speed and a target angle, secondly, adopting a DBSCAN clustering algorithm and a Hampel filtering method, solving noise interference of a dynamic or static target in a range, removing redundant abnormal values at the same time, improving the precision of the human body posture, and therefore constructing a distance-time chart (DTM) and a speed-time chart (VTM), and finally, establishing a multi-dimensional parameter deep learning network framework based on different fusion modes. The network framework uses a convolutional neural network to perform feature extraction on DTM and VTM data sets, uses a low-rank multi-modal fusion (LMF) network to perform feature fusion of different data sets, uses a domain discriminator to obtain features irrelevant to t