Research on Identification Method Based on Point Cloud Sequence with Millimeter-Wave Radar

Authors:

Abstract
The point cloud data obtained from millimeter-wave radar has shown strong potential in pedestrian recognition tasks after tracking a target. This paper proposes a new deep learning model that can extract features from the point cloud sequence data of a target. The proposed model solves the problem of low accuracy of existing methods in pedestrian free-walking scenarios. The model's performance in character recognition and pedestrian recognition was evaluated using both self-constructed and public datasets. The results indicate that the network can efficiently extract features from point cloud sequences, achieving 100% accuracy in recognizing people, fans, and dogs, and 98.9% accuracy in a pedestrian recognition task that includes 20 people, outperforming the current state-of-the-art network.

This publication has 11 references indexed in Scilit: