Human Motion Recognition Method Using Millimeter-wave Radar Based on 3DTSNet
- Resource Type
- Conference
- Authors
- Li, Wei; Wang, Wenxu; Zhang, Duo; Gegentuya
- Source
- 2023 5th International Conference on Electronics and Communication, Network and Computer Technology (ECNCT) Electronics and Communication, Network and Computer Technology (ECNCT), 2023 5th International Conference on. :18-22 Aug, 2023
- Subject
- Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Computing and Processing
Signal Processing and Analysis
Three-dimensional displays
Time series analysis
Radar detection
Millimeter wave radar
Radar imaging
Radar signal processing
Doppler radar
FMCW radar
convolutional neural network
Range Doppler
human action recognition
- Language
This paper proposes a 77 GHz millimeter wave radar-based human motion recognition method. Firstly, a 77 GHz millimeter wave radar data acquisition system is built. Secondly, the radar echoes of multiple human actions are data preprocessed to obtain a 3D Range-Doppler-time image dataset of human actions. Finally, the 3D convolutional neural network model is designed with the human action dataset as the network input sample to complete the simulation experiments of five different human action recognition. The experimental comparison results show that the detection accuracy of the proposed 3DTSNet network model can reach 98.8% and has good generalization.