.png){kind=icon}
Return
Technical Correspondence
.png){kind=icon}
PRE
AI
DOI：10.1109/THMS.2018.2884717.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Human activity recognition techniques based on wearable inertial sensors have achieved great success, but the classification accuracy of human activities using wearable sensors is not good enough in practice. In this paper, a multisensor multiclassifier hierarchical fusion model based on entropy weight for human activity recognition using wearable inertial sensors is proposed. The fusion model has two layers, including basic-classifier fusion layer and sensor fusion layer. The entropy weight method has been applied to achieve the weight values that can affect the decision results of each layer. In addition, a novel feature selection method based on congruent transformation in matrix is also proposed. Three major experiments have been conducted to reveal the feasibility and availability of our algorithms. The experiments show that our fusion algorithm may achieve the better recognition performance when compared with basic classifiers and majority voting. For different feature dimensions, the performance of our algorithm is also better than that of majority voting, and the recognition accuracy rate may reach 96.72%. In addition, the recognition accuracy rate of the proposed feature-selection method is about 96.96%, which is better than the other method.
Keywords:
Body area network
congruent transformation
pattern recognition
sensor network
wearable system
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
4.4
Papers:
1.1K
Citations:
3.5K
Organization
Cited Papers
ShareSaveShareSaveShareSaveShareSaveMulti-sensor fusion in body sensor networks: State-of-the-art and research challenges.png)
INFORMATION FUSION
INFORMATION FUSIONIF15.5
PREAI
ShareSave
View more