Unsupervised Doppler Radar Based Activity Recognition for e-Healthcare

Abstract

Passive radio frequency (RF) sensing and monitoring of human daily activities in elderly care homes is an emerging topic. Micro-Doppler radars are an appealing solution considering their nonintrusiveness, deep penetration, and high-distance range. Unsupervised activity recognition using Doppler radar data has not received attention, in spite of its importance in case of unlabelled or poorly labelled activities in real scenarios. This study proposes two unsupervised feature extraction methods for the purpose of human activity monitoring using Doppler-streams. These include a local Discrete Cosine Transform (DCT)-based feature extraction method and a local entropy-based feature extraction method. In addition, a novel application of Convolutional Variational Autoencoder (CVAE) feature extraction is employed for the first time for Doppler radar data. The three feature extraction architectures are compared with the previously used Convolutional Autoencoder (CAE) and linear feature extraction based on Principal Component Analysis (PCA) and 2DPCA. Unsupervised clustering is performed using K-Means and K-Medoids. The results show the superiority of DCT-based method, entropy-based method, and CVAE features compared to CAE, PCA, and 2DPCA, with more than 5%-20% average accuracy. In regards to computation time, the two proposed methods are noticeably much faster than the existing CVAE. Furthermore, for high-dimensional data visualisation, three manifold learning techniques are considered. The methods are compared for the projection of raw data as well as the encoded CVAE features. All three methods show an improved visualisation ability when applied to the encoded CVAE features.

A Low-Resolution Infrared Array for Unobtrusive Human Activity Recognition That Preserves Privacy

“…At a minimum, a Raspberry Pi processing unit is required to process these data, which takes more time than the previously discussed direct approach. Similar results were obtained with the dataset of the authors of [25]. We tested our method with their dataset and got similar results.…”

Machine Learning for Healthcare Radars: Recent Progresses in Human Vital Sign Measurement and Activity Recognition

“…First, the suitable radar sensor is selected to create a dataset. As per our survey, with few exceptions [111], supervised ML is used which additionally requires dataset-labelling [189]. Furthermore, as expressed in Fig.…”

“…3) Data Analysis and Processing Approaches: Activity recognition with radar has mostly been considered as a classification problem and supervised ML algorithms are being used [27] which additionally requires data labeling. To the best of our knowledge, only two papers used unsupervised learning approach for activity classification using k-means [111] and HMM [110]. Additionally, semi-supervised learning for activity recognition is proposed in references [76], [89], [143].…”

STC-NLSTMNet: An Improved Human Activity Recognition Method Using Convolutional Neural Network with NLSTM from WiFi CSI

“…Karayaneva et al [ 28 ] employed an unsupervised framework based on Doppler radar to recognize daily activities in e-healthcare. They achieved about 80% accuracy.…”