Sparsity-based Human Activity Recognition with PointNet using a Portable FMCW Radar

Journal Article (2023)
Author(s)

Chuanwei Ding (Nanjing University of Science and Technology)

Li Zhang (Nanjing University of Science and Technology)

Haoyu Chen (Nanjing University of Science and Technology)

Hong Hong (Nanjing University of Science and Technology)

Xiaohua Zhu (Nanjing University of Science and Technology)

Francesco Fioranelli (TU Delft - Microwave Sensing, Signals & Systems)

Microwave Sensing, Signals & Systems
Copyright
© 2023 Chuanwei Ding, Li Zhang, Haoyu Chen, Hong Hong, Xiaohua Zhu, F. Fioranelli
DOI related publication
https://doi.org/10.1109/JIOT.2023.3235808
More Info
expand_more
Publication Year
2023
Language
English
Copyright
© 2023 Chuanwei Ding, Li Zhang, Haoyu Chen, Hong Hong, Xiaohua Zhu, F. Fioranelli
Microwave Sensing, Signals & Systems
Issue number
11
Volume number
10
Pages (from-to)
10024-10037
Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Abstract

Radar-based solutions have attracted great attention in human activity recognition (HAR) for their advantages in accuracy, robustness, and privacy protection. The conventional approaches transform radar signals into feature maps and then directly process them as visual images. While effective, these image-based methods may not be the best solutions in terms of representation efficiency to encode the relevant information for classification. This article proposes a novel HAR method combining sparse theory and PointNet network, with both operations in the time-Doppler (TD) and range-Doppler (RD) domains. First, sparsity-based feature extraction is introduced to use a limited number of sparse solutions to characterize human activities in the form of TD sparse point clouds (TDSP) or dynamic RD sparse point clouds (DRDSP). This new representation is validated by comparing the reconstructed and original signals. Then, PointNet networks are adopted to summarize multidomain features and predict human activity labels by a sparse set of input point clouds. Comprehensive experiments were conducted to demonstrate that the proposed method can yield a higher representation efficiency, classification accuracy, and better generalization capability than existing ones.

Files

Sparsity_Based_Human_Activity_... (pdf)
(pdf | 4.06 Mb)
- Embargo expired in 10-07-2023
License info not available