Research Area:  Machine Learning
Sensor-based human activity recognition (HAR) aims to recognize a humans physical actions by using sensors attached to different body parts. As a user-specific application, HAR often suffers poor generalization from training on an individual to testing on another individual, or from one body part to another body part. To tackle this cross-domain HAR problem, this article proposes a domain adaptation (DA) method called local domain adaptation (LDA), whose core is to align cluster-to-cluster distributions between the source domain and the target domain. On the one hand, LDA differs from existing set-to-set alignment by reducing the distribution discrepancy at a finer granularity. On the other hand, LDA is superior to the class-to-class alignment because it can provide more accurate soft labels for the target domain. Specifically, LDA contains three main steps: 1) groups the activity class into several high-level abstract clusters; 2) maps the original data of each cluster in both domains into the same low-dimension subspace to align the intracluster data distribution; 3) predicts the class labels for target domain in the low-dimension subspace. Experimental results on two public HAR benchmark datasets show that LDA outperforms state-of-the-art DA methods for the cross-domain HAR.
Author(s) Name:  Jiachen Zhao; Fang Deng; Haibo He; Jie Chen
Journal name:  IEEE Transactions on Human-Machine Systems
Publisher name:  IEEE
Volume Information:  ( Volume: 51, Issue: 1, Feb. 2021) Page(s): 12 - 21
Paper Link:   https://ieeexplore.ieee.org/abstract/document/9288927