This website requires JavaScript.

Human Activity Recognition from Wi-Fi CSI Data Using Principal Component-Based Wavelet CNN

Ishtiaque Ahmed ShowmikTahsina Farah SanamHafiz Imtiaz
Dec 2022
摘要
Human Activity Recognition (HAR) is an emerging technology with severalapplications in surveillance, security, and healthcare sectors. Noninvasive HARsystems based on Wi-Fi Channel State Information (CSI) signals can be developedleveraging the quick growth of ubiquitous Wi-Fi technologies, and thecorrelation between CSI dynamics and body motions. In this paper, we proposePrincipal Component-based Wavelet Convolutional Neural Network (or PCWCNN) -- anovel approach that offers robustness and efficiency for practical real-timeapplications. Our proposed method incorporates two efficient preprocessingalgorithms -- the Principal Component Analysis (PCA) and the Discrete WaveletTransform (DWT). We employ an adaptive activity segmentation algorithm that isaccurate and computationally light. Additionally, we used the Wavelet CNN forclassification, which is a deep convolutional network analogous to thewell-studied ResNet and DenseNet networks. We empirically show that ourproposed PCWCNN model performs very well on a real dataset, outperformingexisting approaches.
展开全部
图表提取

暂无人提供速读十问回答

论文十问由沈向洋博士提出,鼓励大家带着这十个问题去阅读论文,用有用的信息构建认知模型。写出自己的十问回答,还有机会在当前页面展示哦。

Q1论文试图解决什么问题?
Q2这是否是一个新的问题?
Q3这篇文章要验证一个什么科学假设?